Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

RuntimeError: CUDA error: device-side assert triggered when running Llama on multiple gpus #22546

Closed
2 of 4 tasks
TerryCM opened this issue Apr 3, 2023 · 42 comments
Closed
2 of 4 tasks

RuntimeError: CUDA error: device-side assert triggered when running Llama on multiple gpus #22546

TerryCM opened this issue Apr 3, 2023 · 42 comments

Comments

@TerryCM
Copy link

TerryCM commented Apr 3, 2023 •
edited
Loading

System Info

  • transformers version: 4.28.0.dev0
  • Platform: Linux-3.10.0-1160.81.1.el7.x86_64-x86_64-with-glibc2.17
  • Python version: 3.11.2
  • Huggingface_hub version: 0.13.3
  • Safetensors version: not installed
  • PyTorch version (GPU?): 2.0.0+cu118 (True)
  • Tensorflow version (GPU?): not installed (NA)
  • Flax version (CPU?/GPU?/TPU?): not installed (NA)
  • Jax version: not installed
  • JaxLib version: not installed
  • Using GPU in script?: yes
  • Using distributed or parallel set-up in script?: no

Who can help?

@sgugger @MKhalusova @ArthurZucker @younesbelkada

I am experiencing an assertion error in ScatterGatherKernel.cu when using LlamaTokenizer and multi-GPU inference with any variant of Llama model. The error occurs during the model.generate() call.

import os
# os.environ['TRANSFORMERS_CACHE'] = '/tmp/cache/'
# os.environ['NCCL_P2P_DISABLE'] = '1'

from transformers import AutoModelForCausalLM,AutoConfig,LlamaTokenizer
from accelerate import init_empty_weights, infer_auto_device_map
import torch

def get_device_map(model_path):
    with init_empty_weights():
        config = AutoConfig.from_pretrained(model_path)
        model = AutoModelForCausalLM.from_config(config)

    d = {0: "18GiB"}
    for i in range(1, 5):
        d[i] = "26GiB"
    device_map = infer_auto_device_map(
        model, max_memory=d,dtype=torch.float16, no_split_module_classes=["BloomBlock", "OPTDecoderLayer", "LLaMADecoderLayer", "LlamaDecoderLayer"]
    )
    print(device_map)
    del model
    return device_map




tokenizer = LlamaTokenizer.from_pretrained("decapoda-research/llama-13b-hf")
model = AutoModelForCausalLM.from_pretrained("decapoda-research/llama-13b-hf", torch_dtype=torch.float16, device_map=get_device_map("decapoda-research/llama-13b-hf"))

generate_kwargs = {
    "max_new_tokens": 200,
    "min_new_tokens": 100,
    "temperature": 0.1,
    "do_sample": False, # The three options below used together leads to contrastive search
    "top_k": 4,
    "penalty_alpha": 0.6,
    
    
}


prompt = "Puma is a "
with torch.no_grad():
    input_ids = tokenizer(prompt, return_tensors="pt").input_ids
    assert len(input_ids) == 1, len(input_ids)
    if input_ids[0][-1] == 2: # 2 is EOS, hack to remove. If the prompt is ending with EOS, often the generation will stop abruptly.
        input_ids = input_ids[:, :-1]
    input_ids = input_ids.to(0)
    #input_ids = tokenizer(prompt, padding=True, truncation=True, return_tensors="pt").input_ids.to(0)
    generated_ids = model.generate(
        input_ids,
        #stopping_criteria=stopping_criteria,
        **generate_kwargs
    )
    result = tokenizer.batch_decode(generated_ids.cpu(), skip_special_tokens=True)
    print(result)

The tokenizer class you load from this checkpoint is not the same type as the class this function is called from. It may result in unexpected tokenization.
The tokenizer class you load from this checkpoint is 'LLaMATokenizer'.
The class this function is called from is 'LlamaTokenizer'.
normalizer.cc(51) LOG(INFO) precompiled_charsmap is empty. use identity normalization.
{'model.embed_tokens': 0, 'model.layers.0': 0, 'model.layers.1': 0, 'model.layers.2': 0, 'model.layers.3': 0, 'model.layers.4': 0, 'model.layers.5': 0, 'model.layers.6': 0, 'model.layers.7': 0, 'model.layers.8': 0, 'model.layers.9': 0, 'model.layers.10': 0, 'model.layers.11': 0, 'model.layers.12': 0, 'model.layers.13': 0, 'model.layers.14': 0, 'model.layers.15': 0, 'model.layers.16': 0, 'model.layers.17': 0, 'model.layers.18': 0, 'model.layers.19': 0, 'model.layers.20': 0, 'model.layers.21': 0, 'model.layers.22': 0, 'model.layers.23': 0, 'model.layers.24': 0, 'model.layers.25': 0, 'model.layers.26': 0, 'model.layers.27': 0, 'model.layers.28': 0, 'model.layers.29': 1, 'model.layers.30': 1, 'model.layers.31': 1, 'model.layers.32': 1, 'model.layers.33': 1, 'model.layers.34': 1, 'model.layers.35': 1, 'model.layers.36': 1, 'model.layers.37': 1, 'model.layers.38': 1, 'model.layers.39': 1, 'model.layers.40': 1, 'model.layers.41': 1, 'model.layers.42': 1, 'model.layers.43': 1, 'model.layers.44': 1, 'model.layers.45': 1, 'model.layers.46': 1, 'model.layers.47': 1, 'model.layers.48': 1, 'model.layers.49': 1, 'model.layers.50': 1, 'model.layers.51': 1, 'model.layers.52': 1, 'model.layers.53': 1, 'model.layers.54': 1, 'model.layers.55': 1, 'model.layers.56': 1, 'model.layers.57': 1, 'model.layers.58': 1, 'model.layers.59': 2, 'model.norm': 2, 'lm_head': 2}
Loading checkpoint shards: 100%|██████████████████████████| 61/61 [00:25<00:00, 2.43it/s]
/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/transformers/generation/utils.py:1219: UserWarning: You have modified the pretrained model configuration to control generation. This is a deprecated strategy to control generation and will be removed soon, in a future version. Please use a generation configuration file (see https://huggingface.co/docs/transformers/main_classes/text_generation)
warnings.warn(
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [64,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [65,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [66,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [67,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [68,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [69,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [70,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [71,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [72,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [73,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [74,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [75,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [76,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [77,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [78,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [79,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [80,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [81,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [82,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [83,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [84,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [85,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [86,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [87,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [88,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [89,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [90,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [91,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [92,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [93,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [94,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [95,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [96,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [97,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [98,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [99,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [100,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [101,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [102,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [103,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [104,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [105,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [106,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [107,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [108,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [109,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [110,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [111,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [112,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [113,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [114,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [115,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [116,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [117,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [118,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [119,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [120,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [121,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [122,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [123,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [124,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [125,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [126,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [127,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [0,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [1,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [2,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [3,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [4,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [5,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [6,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [7,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [8,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [9,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [10,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [11,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [12,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [13,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [14,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [15,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [16,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [17,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [18,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [19,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [20,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [21,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [22,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [23,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [24,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [25,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [26,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [27,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [28,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [29,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [30,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [31,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [32,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [33,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [34,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [35,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [36,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [37,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [38,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [39,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [40,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [41,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [42,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [43,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [44,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [45,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [46,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [47,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [48,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [49,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [50,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [51,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [52,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [53,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [54,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [55,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [56,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [57,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [58,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [59,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [60,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [61,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [62,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [63,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [64,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [65,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [66,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [67,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [68,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [69,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [70,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [71,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [72,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [73,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [74,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [75,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [76,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [77,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [78,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [79,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [80,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [81,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [82,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [83,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [84,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [85,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [86,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [87,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [88,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [89,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [90,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [91,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [92,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [93,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [94,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [95,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [96,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [97,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [98,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [99,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [100,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [101,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [102,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [103,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [104,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [105,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [106,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [107,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [108,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [109,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [110,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [111,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [112,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [113,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [114,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [115,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [116,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [117,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [118,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [119,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [120,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [121,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [122,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [123,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [124,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [125,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [126,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [127,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [0,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [1,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [2,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [3,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [4,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [5,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [6,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [7,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [8,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [9,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [10,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [11,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [12,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [13,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [14,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [15,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [16,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [17,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [18,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [19,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [20,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [21,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [22,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [23,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [24,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [25,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [26,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [27,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [28,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [29,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [30,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [31,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [32,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [33,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [34,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [35,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [36,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [37,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [38,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [39,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [40,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [41,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [42,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [43,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [44,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [45,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [46,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [47,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [48,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [49,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [50,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [51,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [52,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [53,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [54,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [55,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [56,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [57,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [58,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [59,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [60,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [61,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [62,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [1,0,0], thread: [63,0,0] Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.
Traceback (most recent call last):
File "/home/u30/terrycruz/chatPaper.py", line 48, in
generated_ids = model.generate(
^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/transformers/generation/utils.py", line 1457, in generate
return self.contrastive_search(
^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/transformers/generation/utils.py", line 1871, in contrastive_search
outputs = self(
^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
output = old_forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 687, in forward
outputs = self.model(
^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 577, in forward
layer_outputs = decoder_layer(
^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
output = old_forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 292, in forward
hidden_states, self_attn_weights, present_key_value = self.self_attn(
^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
output = old_forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/u30/terrycruz/anaconda3/envs/multiple_gpu/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 241, in forward
attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Information

  • The official example scripts
  • My own modified scripts

Tasks

  • An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
  • My own task or dataset (give details below)

Reproduction

  1. run my script: CUDA_LAUNCH_BLOCKING=1 python script.py

Expected behavior

The puma bla bla bla.
@sgugger
Copy link
Collaborator

sgugger commented Apr 3, 2023

This repo is not in sync with the model and tokenizer as implemented in the Transformers library. Sadly, we do not have permission to distribute the weights, so there is no official checkpoint you can use. After you get the official weights from Meta and run the conversion command as documented, you shouldn't have any problem with the model.

@emvw7yf
Copy link

emvw7yf commented Apr 5, 2023 •
edited
Loading

@sgugger I'm experiencing exactly the same error when using official llama weights converted using the huggingface conversion script from the master branch. It happens on the master branch when running inference with accelerate on multiple GPUs (I tried 2x4090 and 4x4090). To reproduce:

p=pipeline("text-generation", "path/to/converted-llama-30b-hf", torch_dtype=torch.float16, device_map="auto")
p("hi there")

This used to work last week. I don't have the exact branch commit ID, but could do git bisect if it'd help.

I'm using pytorch==2.0.0, cuda 11.7, and recent versions of accelearte and bitsandbytes (yes, it also shows the same error with load_in_8bits=True).

@TerryCM TerryCM mentioned this issue Apr 5, 2023
Closed
41 tasks
@sgugger
Copy link
Collaborator

sgugger commented Apr 5, 2023

@emvw7yf Could you print pipeline.model.hf_device_map and report that here? This would help us debug this issue.

@TerryCM
Copy link
Author

TerryCM commented Apr 5, 2023 •
edited
Loading

@sgugger I managed to use the official llama weights and still getting the same error, for llama 7B, using the code from #22546 (comment), and printing pipeline.model.hf_device_map I'm getting the following:

{'model.embed_tokens': 0, 'model.layers.0': 0, 'model.layers.1': 0, 'model.layers.2': 0, 'model.layers.3': 0, 'model.layers.4': 0, 'model.layers.5': 0, 'model.layers.6': 0, 'model.layers.7': 0, 'model.layers.8': 0, 'model.layers.9': 0, 'model.layers.10': 0, 'model.layers.11': 0, 'model.layers.12': 0, 'model.layers.13': 0, 'model.layers.14': 0, 'model.layers.15': 0, 'model.layers.16': 1, 'model.layers.17': 1, 'model.layers.18': 1, 'model.layers.19': 1, 'model.layers.20': 1, 'model.layers.21': 1, 'model.layers.22': 1, 'model.layers.23': 1, 'model.layers.24': 1, 'model.layers.25': 1, 'model.layers.26': 1, 'model.layers.27': 1, 'model.layers.28': 1, 'model.layers.29': 1, 'model.layers.30': 1, 'model.layers.31': 1, 'model.norm': 1, 'lm_head': 1}

@sgugger
Copy link
Collaborator

sgugger commented Apr 5, 2023

So I skimmed through the existing repos to look for one that has the same weights/tokenizer as what I get after the conversion script is applied. Applying this code:

p=pipeline("text-generation", "huggyllama/llama-7b", torch_dtype=torch.float16, device_map="auto")
p("hi there")

gives me the exact same device map as you @TerryCM and works without any issue. I am on Transformers main and Accelerate latest version.

@TerryCM
Copy link
Author

TerryCM commented Apr 5, 2023

@sgugger I'm also on Transformers main and accelerate version (I used pip install accelerate), could be this a drivers problem? Im using the following drivers
Screenshot 2023-04-05 at 10 43 27 AM

@sgugger
Copy link
Collaborator

sgugger commented Apr 5, 2023

Are you using the same repository as me? I'm on CUDA 11.8 and 520 drivers.

@emvw7yf
Copy link

emvw7yf commented Apr 6, 2023 •
edited
Loading

I can reliably reproduce it on both runpod.io and vast.ai. I'm using 2x4090 GPUs and the default docker image on each service (runpod/pytorch:3.10-2.0.0-117 and pytorch/pytorch:2.0.0-cuda11.7-cudnn8-devel).

I'm running the following:

pip install git+https://github.com/huggingface/transformers.git accelerate sentencepiece

import torch
from transformers import pipeline

p=pipeline("text-generation", "huggyllama/llama-7b", torch_dtype=torch.float16, device_map="auto")
p("hi there")

This results in the assertion error above. When I restrict it to a single GPU (using CUDA_VISIBLE_DEVICES), it works without errors.

Versions (taken on vast.ai):

root@C.6113089:/$ nvidia-smi
Wed Apr  5 23:57:43 2023       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.78.01    Driver Version: 525.78.01    CUDA Version: 12.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  NVIDIA GeForce ...  On   | 00000000:41:00.0 Off |                  Off |
|  0%   21C    P8    18W / 450W |      1MiB / 24564MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
|   1  NVIDIA GeForce ...  On   | 00000000:42:00.0 Off |                  Off |
|  0%   21C    P8    24W / 450W |      1MiB / 24564MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
root@C.6113089:/$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Jun__8_16:49:14_PDT_2022
Cuda compilation tools, release 11.7, V11.7.99
Build cuda_11.7.r11.7/compiler.31442593_0
root@C.6113089:/$ python -c "import torch; print(torch.version.cuda)"
11.7

How could I help debugging this?

@emvw7yf
Copy link

emvw7yf commented Apr 6, 2023

I actually realized that the error I'm getting is slightly different (even though the assertion is the same), pasting it below:

/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1313: UserWarning: Using `max_length`'s default (20) to control the generation length. This behaviour is deprecated and will be removed from the config in v5 of Transformers -- we recommend using `max_new_tokens` to control the maximum length of the generation.
  warnings.warn(
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [64,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [65,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [66,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [67,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [68,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [69,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [70,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [71,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [72,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [73,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [74,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [75,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [76,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [77,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [78,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [79,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [80,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [81,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [82,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [83,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [84,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [85,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [86,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [87,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [88,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [89,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [90,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [91,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [92,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [93,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [94,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [95,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [0,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [1,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [2,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [3,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [4,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [5,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [6,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [7,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [8,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [9,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [10,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [11,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [12,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [13,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [14,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [15,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [16,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [17,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [18,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [19,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [20,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [21,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [22,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [23,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [24,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [25,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [26,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [27,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [28,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [29,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [30,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [31,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [96,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [97,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [98,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [99,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [100,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [101,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [102,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [103,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [104,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [105,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [106,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [107,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [108,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [109,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [110,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [111,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [112,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [113,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [114,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [115,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [116,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [117,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [118,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [119,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [120,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [121,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [122,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [123,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [124,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [125,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [126,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [127,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [32,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [33,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [34,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [35,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [36,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [37,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [38,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [39,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [40,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [41,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [42,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [43,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [44,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [45,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [46,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [47,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [48,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [49,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [50,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [51,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [52,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [53,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [54,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [55,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [56,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [57,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [58,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [59,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [60,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [61,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [62,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/opt/conda/conda-bld/pytorch_1678411187366/work/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): block: [0,0,0], thread: [63,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
Cell In[2], line 5
      2 from transformers import pipeline
      4 p=pipeline("text-generation", "huggyllama/llama-7b", torch_dtype=torch.float16, device_map="auto")
----> 5 p("hi there")

File /opt/conda/lib/python3.10/site-packages/transformers/pipelines/text_generation.py:209, in TextGenerationPipeline.__call__(self, text_inputs, **kwargs)
    168 def __call__(self, text_inputs, **kwargs):
    169     """
    170     Complete the prompt(s) given as inputs.
    171 
   (...)
    207           ids of the generated text.
    208     """
--> 209     return super().__call__(text_inputs, **kwargs)

File /opt/conda/lib/python3.10/site-packages/transformers/pipelines/base.py:1109, in Pipeline.__call__(self, inputs, num_workers, batch_size, *args, **kwargs)
   1101     return next(
   1102         iter(
   1103             self.get_iterator(
   (...)
   1106         )
   1107     )
   1108 else:
-> 1109     return self.run_single(inputs, preprocess_params, forward_params, postprocess_params)

File /opt/conda/lib/python3.10/site-packages/transformers/pipelines/base.py:1116, in Pipeline.run_single(self, inputs, preprocess_params, forward_params, postprocess_params)
   1114 def run_single(self, inputs, preprocess_params, forward_params, postprocess_params):
   1115     model_inputs = self.preprocess(inputs, **preprocess_params)
-> 1116     model_outputs = self.forward(model_inputs, **forward_params)
   1117     outputs = self.postprocess(model_outputs, **postprocess_params)
   1118     return outputs

File /opt/conda/lib/python3.10/site-packages/transformers/pipelines/base.py:1015, in Pipeline.forward(self, model_inputs, **forward_params)
   1013     with inference_context():
   1014         model_inputs = self._ensure_tensor_on_device(model_inputs, device=self.device)
-> 1015         model_outputs = self._forward(model_inputs, **forward_params)
   1016         model_outputs = self._ensure_tensor_on_device(model_outputs, device=torch.device("cpu"))
   1017 else:

File /opt/conda/lib/python3.10/site-packages/transformers/pipelines/text_generation.py:251, in TextGenerationPipeline._forward(self, model_inputs, **generate_kwargs)
    249 prompt_text = model_inputs.pop("prompt_text")
    250 # BS x SL
--> 251 generated_sequence = self.model.generate(input_ids=input_ids, attention_mask=attention_mask, **generate_kwargs)
    252 out_b = generated_sequence.shape[0]
    253 if self.framework == "pt":

File /opt/conda/lib/python3.10/site-packages/torch/utils/_contextlib.py:115, in context_decorator.<locals>.decorate_context(*args, **kwargs)
    112 @functools.wraps(func)
    113 def decorate_context(*args, **kwargs):
    114     with ctx_factory():
--> 115         return func(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1437, in GenerationMixin.generate(self, inputs, generation_config, logits_processor, stopping_criteria, prefix_allowed_tokens_fn, synced_gpus, streamer, **kwargs)
   1431         raise ValueError(
   1432             f"num_return_sequences has to be 1, but is {generation_config.num_return_sequences} when doing"
   1433             " greedy search."
   1434         )
   1436     # 11. run greedy search
-> 1437     return self.greedy_search(
   1438         input_ids,
   1439         logits_processor=logits_processor,
   1440         stopping_criteria=stopping_criteria,
   1441         pad_token_id=generation_config.pad_token_id,
   1442         eos_token_id=generation_config.eos_token_id,
   1443         output_scores=generation_config.output_scores,
   1444         return_dict_in_generate=generation_config.return_dict_in_generate,
   1445         synced_gpus=synced_gpus,
   1446         streamer=streamer,
   1447         **model_kwargs,
   1448     )
   1450 elif is_contrastive_search_gen_mode:
   1451     if generation_config.num_return_sequences > 1:

File /opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:2248, in GenerationMixin.greedy_search(self, input_ids, logits_processor, stopping_criteria, max_length, pad_token_id, eos_token_id, output_attentions, output_hidden_states, output_scores, return_dict_in_generate, synced_gpus, streamer, **model_kwargs)
   2245 model_inputs = self.prepare_inputs_for_generation(input_ids, **model_kwargs)
   2247 # forward pass to get next token
-> 2248 outputs = self(
   2249     **model_inputs,
   2250     return_dict=True,
   2251     output_attentions=output_attentions,
   2252     output_hidden_states=output_hidden_states,
   2253 )
   2255 if synced_gpus and this_peer_finished:
   2256     continue  # don't waste resources running the code we don't need

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /opt/conda/lib/python3.10/site-packages/accelerate/hooks.py:165, in add_hook_to_module.<locals>.new_forward(*args, **kwargs)
    163         output = old_forward(*args, **kwargs)
    164 else:
--> 165     output = old_forward(*args, **kwargs)
    166 return module._hf_hook.post_forward(module, output)

File /opt/conda/lib/python3.10/site-packages/transformers/models/llama/modeling_llama.py:687, in LlamaForCausalLM.forward(self, input_ids, attention_mask, position_ids, past_key_values, inputs_embeds, labels, use_cache, output_attentions, output_hidden_states, return_dict)
    684 return_dict = return_dict if return_dict is not None else self.config.use_return_dict
    686 # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
--> 687 outputs = self.model(
    688     input_ids=input_ids,
    689     attention_mask=attention_mask,
    690     position_ids=position_ids,
    691     past_key_values=past_key_values,
    692     inputs_embeds=inputs_embeds,
    693     use_cache=use_cache,
    694     output_attentions=output_attentions,
    695     output_hidden_states=output_hidden_states,
    696     return_dict=return_dict,
    697 )
    699 hidden_states = outputs[0]
    700 logits = self.lm_head(hidden_states)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /opt/conda/lib/python3.10/site-packages/transformers/models/llama/modeling_llama.py:577, in LlamaModel.forward(self, input_ids, attention_mask, position_ids, past_key_values, inputs_embeds, use_cache, output_attentions, output_hidden_states, return_dict)
    569     layer_outputs = torch.utils.checkpoint.checkpoint(
    570         create_custom_forward(decoder_layer),
    571         hidden_states,
   (...)
    574         None,
    575     )
    576 else:
--> 577     layer_outputs = decoder_layer(
    578         hidden_states,
    579         attention_mask=attention_mask,
    580         position_ids=position_ids,
    581         past_key_value=past_key_value,
    582         output_attentions=output_attentions,
    583         use_cache=use_cache,
    584     )
    586 hidden_states = layer_outputs[0]
    588 if use_cache:

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /opt/conda/lib/python3.10/site-packages/accelerate/hooks.py:165, in add_hook_to_module.<locals>.new_forward(*args, **kwargs)
    163         output = old_forward(*args, **kwargs)
    164 else:
--> 165     output = old_forward(*args, **kwargs)
    166 return module._hf_hook.post_forward(module, output)

File /opt/conda/lib/python3.10/site-packages/transformers/models/llama/modeling_llama.py:292, in LlamaDecoderLayer.forward(self, hidden_states, attention_mask, position_ids, past_key_value, output_attentions, use_cache)
    289 hidden_states = self.input_layernorm(hidden_states)
    291 # Self Attention
--> 292 hidden_states, self_attn_weights, present_key_value = self.self_attn(
    293     hidden_states=hidden_states,
    294     attention_mask=attention_mask,
    295     position_ids=position_ids,
    296     past_key_value=past_key_value,
    297     output_attentions=output_attentions,
    298     use_cache=use_cache,
    299 )
    300 hidden_states = residual + hidden_states
    302 # Fully Connected

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /opt/conda/lib/python3.10/site-packages/accelerate/hooks.py:165, in add_hook_to_module.<locals>.new_forward(*args, **kwargs)
    163         output = old_forward(*args, **kwargs)
    164 else:
--> 165     output = old_forward(*args, **kwargs)
    166 return module._hf_hook.post_forward(module, output)

File /opt/conda/lib/python3.10/site-packages/transformers/models/llama/modeling_llama.py:243, in LlamaAttention.forward(self, hidden_states, attention_mask, position_ids, past_key_value, output_attentions, use_cache)
    240 attn_output = attn_output.transpose(1, 2)
    241 attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
--> 243 attn_output = self.o_proj(attn_output)
    245 if not output_attentions:
    246     attn_weights = None

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /opt/conda/lib/python3.10/site-packages/accelerate/hooks.py:165, in add_hook_to_module.<locals>.new_forward(*args, **kwargs)
    163         output = old_forward(*args, **kwargs)
    164 else:
--> 165     output = old_forward(*args, **kwargs)
    166 return module._hf_hook.post_forward(module, output)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/linear.py:114, in Linear.forward(self, input)
    113 def forward(self, input: Tensor) -> Tensor:
--> 114     return F.linear(input, self.weight, self.bias)

RuntimeError: CUDA error: CUBLAS_STATUS_EXECUTION_FAILED when calling `cublasGemmEx( handle, opa, opb, m, n, k, &falpha, a, CUDA_R_16F, lda, b, CUDA_R_16F, ldb, &fbeta, c, CUDA_R_16F, ldc, CUDA_R_32F, CUBLAS_GEMM_DFALT_TENSOR_OP)`

@emvw7yf
Copy link

emvw7yf commented Apr 6, 2023

Interestingly, I'm not getting this error on my home machine. I'm using the same GPUs and the same docker image, so the versions are exactly the same - except the nvidia driver is 525.89.02 instead of 525.78.01.

@kooshi kooshi mentioned this issue Apr 14, 2023
Merged
@github-actions
Copy link

github-actions bot commented May 4, 2023

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

@ghost
Copy link

ghost commented May 21, 2023 •
edited by ghost
Loading

I encountered the same error, with CUDA Version: 11.7 and Driver Version: 515.86.01

@XuehaiPan XuehaiPan mentioned this issue May 21, 2023
Closed
3 tasks
@yuanliping
Copy link

In "config.json" change "pad_token_id=-1" to "pad_token_id=2". This happens because during batch generation, the model sometimes generates pad_token_id=-1

@kevinuserdd
Copy link

I encountered the same error, with CUDA Version: 11.7 and Driver Version: 515.86.01

how to solve?

@PanXiebit
Copy link

pad_token_id

Thanks! This solve my problem.

@RogerYu123
Copy link

same problem

@hjc3613
Copy link

hjc3613 commented Jul 10, 2023

same error when I load model on multiple gpus eg. 4,which set bu CUDA_VISIBLE_DEVICES=0,1,2,3. but when I load model only in 1 gpu, It can generate result succesfully. my code:
`
tokenizer = LlamaTokenizer.from_pretrained(hf_model_path)
model = LlamaForCausalLM.from_pretrained(
hf_model_path,
torch_dtype=torch.float16,
low_cpu_mem_usage=True,
device_map="auto",
load_in_8bit=True
)
generation_output = model.generate(**inputs,
return_dict_in_generate=True,
output_scores=True,
#max_length=512,
max_new_tokens=512,
do_sample=False,
early_stopping=True,
#top_p = 0.6,
num_beams=3,
#eos_token_id=tokenizer.eos_token_id,
num_return_sequences = 1)

sentence = tokenizer.decode(generation_output.sequences[0])

`
how to explain this problem?
transformer version: 4.30.2
accelerate version: 0.20.3

@MoonRainy21
Copy link

MoonRainy21 commented Jul 27, 2023 •
edited
Loading

same error when I load model on multiple gpus

I'm experiencing the same issue with two gpus. When I replace device_map="auto" to device_map={"":"cuda:0"} the model generates as expected.
I'm using two A6000s.
CUDA Version: 12.2
CUDA Driver: 535.54.03
transformer version: 4.28.1
accelerate version: 0.20.3
Python: 3.8.10
torch: 2.0.1

@wang-sj16
Copy link

same problem when running with multiple gpus

@xhran2010
Copy link

same problem here

@sgugger
Copy link
Collaborator

sgugger commented Aug 2, 2023

Please stop commenting with "same problem" without providing a reproducer. We can't do anything about a bug we can't reproduce.

@wang-sj16
Copy link

@sgugger sorry, here's my environment:
Two A6000s.
CUDA Version: 11.7
transformer version: 4.32.0.dev
accelerate version: 0.21.0
Python: 3.9.16
torch: 2.0.1

@fake-warrior8
Copy link

In "config.json" change "pad_token_id=-1" to "pad_token_id=2". This happens because during batch generation, the model sometimes generates pad_token_id=-1

why set pad_token_id to 2 instead of 0? Does this (set pad_token_id to 2) have any impact on the model performance?

@thusithaC
Copy link

I could reproduce the error with following env.

2x A100 (80 GB each)

python 3.10.6
torch==2.0.0
transformers[torch]==4.31.0
accelerate==0.21.0
Driver Version: 535.54.03

torch.version.cuda -> 11.8

Model "meta-llama/Llama-2-70b-chat-hf"

tokenizer = AutoTokenizer.from_pretrained(model_name)
        model = AutoModelForCausalLM.from_pretrained(
            model_name,
            torch_dtype=torch.bfloat16,
            trust_remote_code=True,
            load_in_8bit=False,
            device_map="auto",
        )
        pipeline_ = pipeline(
            "text-generation",
            model=model,
            tokenizer=tokenizer,
        )

Worked fine with 1xA100 with 8 bit model set to true. I just wanted to run some test with 16 bit mode.

@shl518
Copy link

shl518 commented Aug 10, 2023 •
edited
Loading

I could reproduce the error with following env.

2x A6000 (48 GB each)

python 3.8
torch==1.13.1
transformers[torch]==4.31.0
accelerate==0.21.0
Driver Version: 535.54.03

torch.version.cuda -> 11.7

Model "meta-llama/Llama-2-7b-chat-hf"

Following is part of my code for loading model:

tokenizer = AutoTokenizer.from_pretrained(model_name)
kwargs = {"device_map":"balanced_low_0","torch_dtype":torch.float16}
model = AutoModelForCausalLM.from_pretrained(model_name, **kwargs)

Following is part of my code for process data:

# make llama 2 chat prompt
input_sentences = [
    "DeepSpeed is a machine learning framework",
    "He is working on",
    "He has a",
    "He got all",
    "Everyone is happy and I can",
    "The new movie that got Oscar this year",
    "In the far far distance from our galaxy,",
    "Peace is the only way",
]
inputs = input_sentences[: args.batch_size]
system_message = "You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe.  Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature. If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information."

prompt_template = f'''[INST] <<SYS>>
{system_message}
<</SYS>>

{{prompt}} [/INST]
'''
processed_prompts = [prompt_template.format(prompt=p) for p in inputs]

Following is part of my code for generate:

def generate():
    """returns a list of zipped inputs, outputs and number of new tokens"""
    input_tokens = tokenizer.batch_encode_plus(
        processed_prompts, return_tensors="pt", padding=False)
    for t in input_tokens:
        if torch.is_tensor(input_tokens[t]):
            input_tokens[t] = input_tokens[t].to("cuda:0")

    outputs = model.generate(**input_tokens, **generate_kwargs)
    input_tokens_lengths = [x.shape[0] for x in input_tokens.input_ids]
    output_tokens_lengths = [x.shape[0] for x in outputs]

    total_new_tokens = [o - i for i,
                        o in zip(input_tokens_lengths, output_tokens_lengths)]
    outputs = tokenizer.batch_decode(
        outputs, skip_special_tokens=True, spaces_between_special_tokens=False)

    return zip(inputs, outputs, total_new_tokens)

This error occurs only when I am running on multiple Gpus, and when I debug it turns out that the model inference results are obviously wrong，like this
截屏2023-08-10 11 21 32
After checking my model weights，I'm sure it has no problem. I also try to covert Llama-2-7b-chat to Llama-2-7b-chat-hf by myself, but the error still occur.Hope someone can help me solve this problem

@sgugger
Copy link
Collaborator

sgugger commented Aug 10, 2023

@thusithaC Does it only happen for the 70b model or does it also happen with the 7b model? cc @SunMarc

@shl518 This is not a reproducer, it lacks how the model is created or what prompts you pass it.

@shl518
Copy link

shl518 commented Aug 10, 2023 •
edited
Loading

@thusithaC Does it only happen for the 70b model or does it also happen with the 7b model? cc @SunMarc

@shl518 This is not a reproducer, it lacks how the model is created or what prompts you pass it.

sorry, I have update my issue

@thusithaC
Copy link

hi @sgugger I could only reproduce it for the 70B model. The trigger condition seemed to be getting split up in multiple GPUs, and was difficult to do so with the 13/7 B models.

@takitsuba takitsuba mentioned this issue Aug 20, 2023
Closed
@takitsuba
Copy link

I had a similar error.
In my case, the cause is the late communication speed between GPUs.( I checked p2pBandwidthLatencyTest )
I solved it by changing iommu setting like this.

reference

@kerenganon
Copy link

This is happening for me too:

I'm running on 2 24G A5000 GPUs:
image

This my code for loading the model:

    def __init__(self):
        model_name = "meta-llama/Llama-2-70b-chat-hf"
        tokenizer = AutoTokenizer.from_pretrained(model_name)
        tokenizer.pad_token_id = tokenizer.eos_token_id
        tokenizer.pad_token = "[PAD]"
        tokenizer.padding_side = "left"

        bnb_config = BitsAndBytesConfig(
            load_in_4bit=True,
            bnb_4bit_quant_type="nf4",
            bnb_4bit_compute_dtype=torch.float16,
            bnb_4bit_use_double_quant=True
        )
        model = AutoModelForCausalLM.from_pretrained(
            model_name,
            quantization_config=bnb_config,
            device_map="auto",
            trust_remote_code=True,
        )

        pipe = pipeline(
            "text-generation",
            model=model,
            tokenizer=tokenizer,
            trust_remote_code=True,
            device_map="auto"
        )

        self.tokenizer = tokenizer
        self.model = model
        self.pipe = pipe

When I try prompting it it crashes with this error:

RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
Full trace: 
Traceback (most recent call last):
  File "/root/miniconda3/envs/floorplans/lib/python3.11/runpy.py", line 198, in _run_module_as_main
    return _run_code(code, main_globals, None,
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/runpy.py", line 88, in _run_code
    exec(code, run_globals)
  File "/root/.vscode-server/extensions/ms-python.python-2023.16.0/pythonFiles/lib/python/debugpy/adapter/../../debugpy/launcher/../../debugpy/__main__.py", line 39, in <module>
    cli.main()
  File "/root/.vscode-server/extensions/ms-python.python-2023.16.0/pythonFiles/lib/python/debugpy/adapter/../../debugpy/launcher/../../debugpy/../debugpy/server/cli.py", line 430, in main
    run()
  File "/root/.vscode-server/extensions/ms-python.python-2023.16.0/pythonFiles/lib/python/debugpy/adapter/../../debugpy/launcher/../../debugpy/../debugpy/server/cli.py", line 284, in run_file
    runpy.run_path(target, run_name="__main__")
  File "/root/.vscode-server/extensions/ms-python.python-2023.16.0/pythonFiles/lib/python/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path
    return _run_module_code(code, init_globals, run_name,
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/.vscode-server/extensions/ms-python.python-2023.16.0/pythonFiles/lib/python/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code
    _run_code(code, mod_globals, init_globals,
  File "/root/.vscode-server/extensions/ms-python.python-2023.16.0/pythonFiles/lib/python/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code
    exec(code, run_globals)
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/extract_metadata_info.py", line 232, in <module>
    main()
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/extract_metadata_info.py", line 202, in main
    add_column_to_df(
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/extract_metadata_info.py", line 164, in add_column_to_df
    df[col_name] = iterable_df.progress_apply(
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/tqdm/std.py", line 920, in inner
    return getattr(df, df_function)(wrapper, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/pandas/core/frame.py", line 9568, in apply
    return op.apply().__finalize__(self, method="apply")
           ^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/pandas/core/apply.py", line 764, in apply
    return self.apply_standard()
           ^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/pandas/core/apply.py", line 891, in apply_standard
    results, res_index = self.apply_series_generator()
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/pandas/core/apply.py", line 907, in apply_series_generator
    results[i] = self.f(v)
                 ^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/tqdm/std.py", line 915, in wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/extract_metadata_info.py", line 165, in <lambda>
    lambda row: col_func(llm, prompt, row),
                ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/extract_metadata_info.py", line 139, in get_legend_existence_answer
    ans = get_single_token_answer(llm, prompt)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/extract_metadata_info.py", line 93, in get_single_token_answer
    ans_raw = llm.run_prompt(prompt, max_new_tokens=1)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/llm.py", line 46, in run_prompt
    out = self.pipe(p, **kwargs)
          ^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/text_generation.py", line 205, in __call__
    return super().__call__(text_inputs, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/base.py", line 1140, in __call__
    return self.run_single(inputs, preprocess_params, forward_params, postprocess_params)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/base.py", line 1147, in run_single
    model_outputs = self.forward(model_inputs, **forward_params)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/base.py", line 1046, in forward
    model_outputs = self._forward(model_inputs, **forward_params)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/text_generation.py", line 268, in _forward
    generated_sequence = self.model.generate(input_ids=input_ids, attention_mask=attention_mask, **generate_kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/generation/utils.py", line 1648, in generate
    return self.sample(
           ^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/generation/utils.py", line 2730, in sample
    outputs = self(
              ^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
    output = old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 820, in forward
    outputs = self.model(
              ^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 708, in forward
    layer_outputs = decoder_layer(
                    ^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
    output = old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 424, in forward
    hidden_states, self_attn_weights, present_key_value = self.self_attn(
                                                          ^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
    output = old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 333, in forward
    query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)
                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 187, in apply_rotary_pos_emb
    k_embed = (k * cos) + (rotate_half(k) * sin)
                           ~~~~~~~~~~~~~~~^~~~~
RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

pipe.model.hf_device_map returns:

{'model.embed_tokens': 0, 'model.layers.0': 0, 'model.layers.1': 0, 'model.layers.2': 0, 'model.layers.3': 0, 'model.layers.4': 0, 
'model.layers.5': 0, 'model.layers.6': 0, 'model.layers.7': 0, 'model.layers.8': 0, 'model.layers.9': 0, 'model.layers.10': 0, 
'model.layers.11': 0, 'model.layers.12': 0, 'model.layers.13': 0, 'model.layers.14': 0, 'model.layers.15': 0, 'model.layers.16': 0, 
'model.layers.17': 0, 'model.layers.18': 0, 'model.layers.19': 0, 'model.layers.20': 0, 'model.layers.21': 0, 'model.layers.22': 0, 
'model.layers.23': 0, 'model.layers.24': 0, 'model.layers.25': 0, 'model.layers.26': 0, 'model.layers.27': 0, 'model.layers.28': 0, 
'model.layers.29': 0, 'model.layers.30': 0, 'model.layers.31': 0, 'model.layers.32': 0, 'model.layers.33': 0, 'model.layers.34': 0, 
'model.layers.35': 0, 'model.layers.36': 1, 'model.layers.37': 1, 'model.layers.38': 1, 'model.layers.39': 1, 'model.layers.40': 1, 
'model.layers.41': 1, 'model.layers.42': 1, 'model.layers.43': 1, 'model.layers.44': 1, 'model.layers.45': 1, 'model.layers.46': 1, 
'model.layers.47': 1, 'model.layers.48': 1, 'model.layers.49': 1, 'model.layers.50': 1, 'model.layers.51': 1, 'model.layers.52': 1, 
'model.layers.53': 1, 'model.layers.54': 1, 'model.layers.55': 1, 'model.layers.56': 1, 'model.layers.57': 1, 'model.layers.58': 1, 
'model.layers.59': 1, 'model.layers.60': 1, 'model.layers.61': 1, 'model.layers.62': 1, 'model.layers.63': 1, 'model.layers.64': 1, 
'model.layers.65': 1, 'model.layers.66': 1, 'model.layers.67': 1, 'model.layers.68': 1, 'model.layers.69': 1, 'model.layers.70': 1, 
'model.layers.71': 1, 'model.layers.72': 1, 'model.layers.73': 1, 'model.layers.74': 1, 'model.layers.75': 1, 'model.layers.76': 1, 
'model.layers.77': 1, 'model.layers.78': 1, 'model.layers.79': 1, 'model.norm': 1, 'lm_head': 1}

I've tried running the 13B model on 2 GPUs and got a slightly different behaviour.

I was able to prompt it for a few times and it returned random words, and then crashed with the same error, but from a different line:

  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 346, in forward
    attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
Full trace: 
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/storage/kerenganon/floor_plans/dataset_creation/llms/llm.py", line 46, in run_prompt
    out = self.pipe(p, **kwargs)
          ^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/text_generation.py", line 205, in __call__
    return super().__call__(text_inputs, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/base.py", line 1140, in __call__
    return self.run_single(inputs, preprocess_params, forward_params, postprocess_params)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/base.py", line 1147, in run_single
    model_outputs = self.forward(model_inputs, **forward_params)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/base.py", line 1046, in forward
    model_outputs = self._forward(model_inputs, **forward_params)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/pipelines/text_generation.py", line 268, in _forward
    generated_sequence = self.model.generate(input_ids=input_ids, attention_mask=attention_mask, **generate_kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/generation/utils.py", line 1648, in generate
    return self.sample(
           ^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/generation/utils.py", line 2730, in sample
    outputs = self(
              ^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
    output = old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 820, in forward
    outputs = self.model(
              ^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 708, in forward
    layer_outputs = decoder_layer(
                    ^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
    output = old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 424, in forward
    hidden_states, self_attn_weights, present_key_value = self.self_attn(
                                                          ^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/accelerate/hooks.py", line 165, in new_forward
    output = old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/root/miniconda3/envs/floorplans/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py", line 346, in forward
    attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
                   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~
RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

I've seen this error too:

Error device-side assert triggered at line 88 in file /mmfs1/gscratch/zlab/timdettmers/git/bitsandbytes/csrc/ops.cu

I've managed to run the 13B model on 1 GPU with this code successfully, so I assume there is an issue with the GPU communication?
Any help would be appreciated🙏

@kerenganon kerenganon mentioned this issue Oct 2, 2023
Closed
4 tasks
This was referenced Oct 7, 2023
Closed
Open
@ShayDuane
Copy link

In "config.json" change "pad_token_id=-1" to "pad_token_id=2". This happens because during batch generation, the model sometimes generates pad_token_id=-1

tokenizer.pad_token = tokenizer.eos_token
successful!
but, why?

@RonanKMcGovern
Copy link

RonanKMcGovern commented Nov 2, 2023 •
edited
Loading

I'm getting a similar issue with 2x A100s for inference (or training too). [Although I don't get this issue with 4x A6000s]

Environment

- `transformers` version: 4.36.0.dev0
- Platform: Linux-5.4.0-149-generic-x86_64-with-glibc2.35
- Python version: 3.10.6
- Huggingface_hub version: 0.18.0
- Safetensors version: 0.4.0
- Accelerate version: 0.25.0.dev0
- Accelerate config: 	not found
- PyTorch version (GPU?): 2.0.1+cu118 (True)
- Tensorflow version (GPU?): not installed (NA)
- Flax version (CPU?/GPU?/TPU?): not installed (NA)
- Jax version: not installed
- JaxLib version: not installed
- Using GPU in script?: Yes, 2x A100s.
- Using distributed or parallel set-up in script?: Model split across gpus.

Reproduction:

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_id  = "tiiuae/falcon-40B"

model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map={"":"cuda:0"}, # I also tried 'auto'
    # trust_remote_code=False,
    torch_dtype=torch.bfloat16,
    use_flash_attention_2=True, # works with Llama models and reduces memory reqs)

import torch
import gc

tokenizer = AutoTokenizer.from_pretrained(model_id,use_fast=True)

def generate_simple(model, tokenizer, prompt):
    inputs = tokenizer.encode(prompt, return_tensors="pt").to("cuda")
    outputs = model.generate(inputs, max_length=50)
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

# Basic test of model generation
prompt = "The quick brown fox"
generated_text = generate_simple(model, tokenizer, prompt)
print(generated_text)

Error

The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.
Setting `pad_token_id` to `eos_token_id`:11 for open-end generation.
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
Cell In[9], line 13
     11 # Basic test of model generation
     12 prompt = "The quick brown fox"
---> 13 generated_text = generate_simple(model, tokenizer, prompt)
     14 print(generated_text)
     16 # Clear GPU cache and run garbage collection

Cell In[9], line 8, in generate_simple(model, tokenizer, prompt)
      6 def generate_simple(model, tokenizer, prompt):
      7     inputs = tokenizer.encode(prompt, return_tensors="pt").to("cuda")
----> 8     outputs = model.generate(inputs, max_length=50)
      9     return tokenizer.decode(outputs[0], skip_special_tokens=True)

File /usr/local/lib/python3.10/dist-packages/torch/utils/_contextlib.py:115, in context_decorator.<locals>.decorate_context(*args, **kwargs)
    112 @functools.wraps(func)
    113 def decorate_context(*args, **kwargs):
    114     with ctx_factory():
--> 115         return func(*args, **kwargs)

File /usr/local/lib/python3.10/dist-packages/transformers/generation/utils.py:1753, in GenerationMixin.generate(self, inputs, generation_config, logits_processor, stopping_criteria, prefix_allowed_tokens_fn, synced_gpus, assistant_model, streamer, negative_prompt_ids, negative_prompt_attention_mask, **kwargs)
   1736     return self.assisted_decoding(
   1737         input_ids,
   1738         assistant_model=assistant_model,
   (...)
   1749         **model_kwargs,
   1750     )
   1751 if generation_mode == GenerationMode.GREEDY_SEARCH:
   1752     # 11. run greedy search
-> 1753     return self.greedy_search(
   1754         input_ids,
   1755         logits_processor=logits_processor,
   1756         stopping_criteria=stopping_criteria,
   1757         pad_token_id=generation_config.pad_token_id,
   1758         eos_token_id=generation_config.eos_token_id,
   1759         output_scores=generation_config.output_scores,
   1760         return_dict_in_generate=generation_config.return_dict_in_generate,
   1761         synced_gpus=synced_gpus,
   1762         streamer=streamer,
   1763         **model_kwargs,
   1764     )
   1766 elif generation_mode == GenerationMode.CONTRASTIVE_SEARCH:
   1767     if not model_kwargs["use_cache"]:

File /usr/local/lib/python3.10/dist-packages/transformers/generation/utils.py:2614, in GenerationMixin.greedy_search(self, input_ids, logits_processor, stopping_criteria, max_length, pad_token_id, eos_token_id, output_attentions, output_hidden_states, output_scores, return_dict_in_generate, synced_gpus, streamer, **model_kwargs)
   2611 model_inputs = self.prepare_inputs_for_generation(input_ids, **model_kwargs)
   2613 # forward pass to get next token
-> 2614 outputs = self(
   2615     **model_inputs,
   2616     return_dict=True,
   2617     output_attentions=output_attentions,
   2618     output_hidden_states=output_hidden_states,
   2619 )
   2621 if synced_gpus and this_peer_finished:
   2622     continue  # don't waste resources running the code we don't need

File /usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /usr/local/lib/python3.10/dist-packages/transformers/models/falcon/modeling_falcon.py:1246, in FalconForCausalLM.forward(self, input_ids, past_key_values, attention_mask, position_ids, head_mask, inputs_embeds, labels, use_cache, output_attentions, output_hidden_states, return_dict)
   1237 r"""
   1238 labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
   1239     Labels for language modeling. Note that the labels **are shifted** inside the model, i.e. you can set
   1240     `labels = input_ids` Indices are selected in `[-100, 0, ..., config.vocab_size]` All labels set to `-100`
   1241     are ignored (masked), the loss is only computed for labels in `[0, ..., config.vocab_size]`
   1242 """
   1244 return_dict = return_dict if return_dict is not None else self.config.use_return_dict
-> 1246 transformer_outputs = self.transformer(
   1247     input_ids,
   1248     past_key_values=past_key_values,
   1249     attention_mask=attention_mask,
   1250     position_ids=position_ids,
   1251     head_mask=head_mask,
   1252     inputs_embeds=inputs_embeds,
   1253     use_cache=use_cache,
   1254     output_attentions=output_attentions,
   1255     output_hidden_states=output_hidden_states,
   1256     return_dict=return_dict,
   1257 )
   1258 hidden_states = transformer_outputs[0]
   1260 lm_logits = self.lm_head(hidden_states)

File /usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /usr/local/lib/python3.10/dist-packages/transformers/models/falcon/modeling_falcon.py:1122, in FalconModel.forward(self, input_ids, past_key_values, attention_mask, position_ids, head_mask, inputs_embeds, use_cache, output_attentions, output_hidden_states, return_dict)
   1110     outputs = self._gradient_checkpointing_func(
   1111         block.__call__,
   1112         hidden_states,
   (...)
   1119         output_attentions,
   1120     )
   1121 else:
-> 1122     outputs = block(
   1123         hidden_states,
   1124         layer_past=layer_past,
   1125         attention_mask=attention_mask,
   1126         position_ids=position_ids,
   1127         head_mask=head_mask[i],
   1128         use_cache=use_cache,
   1129         output_attentions=output_attentions,
   1130         alibi=alibi,
   1131     )
   1133 hidden_states = outputs[0]
   1134 if use_cache is True:

File /usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /usr/local/lib/python3.10/dist-packages/transformers/models/falcon/modeling_falcon.py:796, in FalconDecoderLayer.forward(self, hidden_states, alibi, attention_mask, position_ids, layer_past, head_mask, use_cache, output_attentions, **kwargs)
    793     attention_layernorm_out = self.input_layernorm(hidden_states)
    795 # Self attention.
--> 796 attn_outputs = self.self_attention(
    797     attention_layernorm_out,
    798     layer_past=layer_past,
    799     attention_mask=attention_mask,
    800     position_ids=position_ids,
    801     alibi=alibi,
    802     head_mask=head_mask,
    803     use_cache=use_cache,
    804     output_attentions=output_attentions,
    805     **kwargs,
    806 )
    808 attention_output = attn_outputs[0]
    810 if not self.config.new_decoder_architecture:

File /usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /usr/local/lib/python3.10/dist-packages/transformers/models/falcon/modeling_falcon.py:559, in FalconFlashAttention2.forward(self, hidden_states, alibi, attention_mask, position_ids, layer_past, head_mask, use_cache, output_attentions, **kwargs)
    556     # overwrite attention_mask with padding_mask
    557     attention_mask = kwargs.pop("padding_mask")
--> 559 fused_qkv = self.query_key_value(hidden_states)  # [batch_size, seq_length, 3 x hidden_size]
    560 num_kv_heads = self.num_heads if self.new_decoder_architecture else self.num_kv_heads
    561 # 3 x [batch_size, seq_length, num_heads, head_dim]

File /usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501, in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File /usr/local/lib/python3.10/dist-packages/transformers/models/falcon/modeling_falcon.py:68, in FalconLinear.forward(self, input)
     67 def forward(self, input: torch.Tensor) -> torch.Tensor:
---> 68     hidden_states = input @ self.weight.T
     69     if self.bias is None:
     70         return hidden_states

RuntimeError: CUDA error: CUBLAS_STATUS_NOT_INITIALIZED when calling `cublasCreate(handle)`

@hriaz17
Copy link

hriaz17 commented Dec 1, 2023

Any permanent solution to this? Getting the same error fine-tuning meta-llama/Llama-2-70b-hf on 2x A100.

@ArthurZucker
Copy link
Collaborator

This is an issue with the way cuda is installed, not transformers

@ShoufaChen
Copy link

Just in case it's helpful, I encounted this similar error and I found that I forgot to change model's embedding size after adding special tokens (add_special_tokens).

tokenizer.add_special_tokens(...)

# !!! resize after adding tokens !!!
model.resize_token_embeddings(len(tokenizer))

Similar to #24698.

@RonanKMcGovern
Copy link

This is an issue with the way cuda is installed, not transformers

Have you fix or is there a cuda/nvidia repo where it's best to post? Thanks

@abcbdf
Copy link

abcbdf commented Dec 28, 2023

I finally solved this by disabling ACS in bios, ref https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/troubleshooting.html#pci-access-control-services-acs. Changing nvidia driver and cuda version doesn't help.

This test is very helpful. https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/troubleshooting.html#gpu-to-gpu-communication

@ttkrpink
Copy link

My code worked when I had two GPUs. I added one more GPU and updated my driver yesterday. It breaks my code. Error code:
Error device-side assert triggered at line 88 in file /mmfs1/gscratch/zlab/timdettmers/git/bitsandbytes/csrc/ops.cu

As for the ACS, it's disabled.

@asherisaac
Copy link

solved for me by resizing token embeddings after adding pad token.

@rohithbojja
Copy link

rohithbojja commented Mar 19, 2024 •
edited
Loading

@asherisaac
what to resize, i got this ,
#print(tokenizer.all_special_tokens) #-->[ ['', '', '', ''] ]
#print(tokenizer.all_special_ids) #--> [1, 2, 0, 0]
im trying to run my finetuned llava
this is the error:
../aten/src/ATen/native/cuda/Indexing.cu:1146: indexSelectLargeIndex: block: [372,0,0], thread: [31,0,0] Assertion srcIndex < srcSelectDimSize failed.

@shun1267
Copy link

shun1267 commented Apr 4, 2024

With a setup consisting of 4x RTX 3090 GPUs and a TRX40 motherboard. I run my model in a Docker container and encountered a similar CUDA error.

The issue was solved by updating the NVIDIA driver to the latest version and upgrading CUDA from 12.3 to 12.4 on the host machine, without changing the CUDA version within the container. Hope the solution be helpful for those experiencing similar issues in containerized environments.

Note: My host OS is Arch Linux with kernel version 6.8.2-arch2-1, and both CUDA 11.7 and CUDA 12.4 work in the container.

@benjaminye benjaminye mentioned this issue May 21, 2024
Closed
@zzxslp
Copy link

zzxslp commented Jul 18, 2024

This is an issue with the way cuda is installed, not transformers

Hi, could you please explain a bit more why it is an issue with CUDA, and how do we solve it?

@jmamou jmamou mentioned this issue Dec 2, 2024
Open
6 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests