[Badcase]: 使用llamafactory量化gptq时，使用2个GPU时出错

[czhcc]

Model Series

Qwen2.5

What are the models used?

qwen2.5-7b

What is the scenario where the problem happened?

使用llamafactory将qwen2.5-7b量化为gptq

Is this badcase known and can it be solved using avaiable techniques?

• I have followed the GitHub README.
• I have checked the Qwen documentation and cannot find a solution there.
• I have checked the documentation of the related framework and cannot find useful information.
• I have searched the issues and there is not a similar one.

Information about environment

llamafactory version: 0.9.0
Platform: Linux-5.11.0-46-generic-x86_64-with-glibc2.35
Python version: 3.10.12
PyTorch version: 2.5.1+cu124 (GPU)
Transformers version: 4.44.2
Datasets version: 2.21.0
Accelerate version: 0.34.2
PEFT version: 0.12.0
TRL version: 0.9.6
GPU type: NVIDIA A800 80GB PCIe
DeepSpeed version: 0.14.4
Bitsandbytes version: 0.44.1

Description

如果设置CUDA_VISIBLE_DEVICES=0，执行量化没问题，但如果设置CUDA_VISIBLE_DEVICES=0,1时，会出异常
Quantizing model.layers blocks : 48%|████████████████████▎ | 31/64 [21:46<23:11, 42.16s/it]Traceback (most recent call last):
File "/usr/local/bin/llamafactory-cli", line 8, in
sys.exit(main())
File "/app/src/llamafactory/cli.py", line 87, in main
export_model()
File "/app/src/llamafactory/train/tuner.py", line 76, in export_model
model = load_model(tokenizer, model_args, finetuning_args) # must after fixing tokenizer to resize vocab
File "/app/src/llamafactory/model/loader.py", line 166, in load_model
model = load_class.from_pretrained(**init_kwargs)
File "/usr/local/lib/python3.10/dist-packages/transformers/models/auto/auto_factory.py", line 564, in from_pretrained
return model_class.from_pretrained(
File "/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py", line 4037, in from_pretrained
hf_quantizer.postprocess_model(model)
File "/usr/local/lib/python3.10/dist-packages/transformers/quantizers/base.py", line 195, in postprocess_model
return self._process_model_after_weight_loading(model, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/transformers/quantizers/quantizer_gptq.py", line 85, in _process_model_after_weight_loading
self.optimum_quantizer.quantize_model(model, self.quantization_config.tokenizer)
File "/usr/local/lib/python3.10/dist-packages/torch/utils/_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/optimum/gptq/quantizer.py", line 512, in quantize_model
block(*layer_inputs[j], **layer_input_kwargs[j])
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1736, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1747, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/accelerate/hooks.py", line 170, in new_forward
output = module._old_forward(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 655, in forward
hidden_states, self_attn_weights, present_key_value = self.self_attn(
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1736, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1747, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/accelerate/hooks.py", line 170, in new_forward
output = module._old_forward(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 555, in forward
query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)
File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 206, in apply_rotary_pos_emb
q_embed = (q * cos) + (rotate_half(q) * sin)
File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 179, in rotate_half
return torch.cat((-x2, x1), dim=-1)
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.

[jklj077]

could be similar to this huggingface/optimum#1889

[github-actions]

This issue has been automatically marked as inactive due to lack of recent activity. Should you believe it remains unresolved and warrants attention, kindly leave a comment on this thread.