log GPU memory usage

requirements.txt

@@ -19,3 +19,4 @@ evaluate==0.4.0
 rouge-score==0.1.2
 scipy
 scikit-learn==1.2.2
+nvidia-ml-py3

scripts/finetune.py

@@ -18,6 +18,7 @@
 from transformers import GenerationConfig, TextStreamer
 
 from axolotl.logging_config import configure_logging
+from axolotl.utils.bench import log_gpu_memory_usage
 from axolotl.utils.data import load_prepare_datasets, load_pretraining_dataset
 from axolotl.utils.dict import DictDefault
 from axolotl.utils.models import load_model, load_tokenizer
@@ -250,6 +251,8 @@ def train(
         LOG.info("Finished preparing dataset. Exiting...")
         return
 
+    log_gpu_memory_usage(LOG, "baseline", cfg.device)
+
     # Load the model and tokenizer
     LOG.info("loading model and peft_config...")
     model, peft_config = load_model(

src/axolotl/utils/bench.py

@@ -0,0 +1,23 @@
+"""Benchmarking and measurement utilities"""
+
+import pynvml
+import torch
+
+
+def gpu_memory_usage(device):
+    if isinstance(device, torch.device):
+        device = device.index
+    if isinstance(device, str) and device.startswith("cuda:"):
+        device = int(device[5:])
+
+    # NB torch.cuda.memory_usage returns zero so we use lower level api
+    pynvml.nvmlInit()
+    handle = pynvml.nvmlDeviceGetHandleByIndex(device)
+    info = pynvml.nvmlDeviceGetMemoryInfo(handle)
+    return info.used / 1024.0**3
+
+
+def log_gpu_memory_usage(log, msg, device):
+    log.info(
+        f"GPU memory usage {msg}: {gpu_memory_usage(device):.03f} GB", stacklevel=2
+    )

src/axolotl/utils/callbacks.py

@@ -1,5 +1,6 @@
 """Callbacks for Trainer class"""
 
+import logging
 import os
 
 from optimum.bettertransformer import BetterTransformer
@@ -11,6 +12,10 @@
 )
 from transformers.trainer_utils import PREFIX_CHECKPOINT_DIR, IntervalStrategy
 
+from axolotl.utils.bench import log_gpu_memory_usage
+
+LOG = logging.getLogger("axolotl.callbacks")
+
 
 class SavePeftModelCallback(TrainerCallback):  # pylint: disable=too-few-public-methods
     """Callback to save the PEFT adapter"""
@@ -67,3 +72,25 @@ def on_step_end(
             # the trainer will raise an exception since it can't save a BetterTransformer wrapped model
             control.should_save = False
         return control
+
+
+class PrintGPUStatsCallback(
+    TrainerCallback
+):  # pylint: disable=too-few-public-methods disable=unused-argument
+    """Callback to print GPU utilization"""
+
+    def __init__(self, cfg):
+        self.cfg = cfg
+        self.logged = False
+
+    def on_step_end(
+        self,
+        args: TrainingArguments,
+        state: TrainerState,
+        control: TrainerControl,
+        **kwargs,
+    ):
+        if not self.logged:
+            log_gpu_memory_usage(LOG, "while training", self.cfg.device)
+            self.logged = True
+        return control

src/axolotl/utils/models.py

@@ -22,6 +22,7 @@
 )
 
 from axolotl.prompt_tokenizers import LLAMA_DEFAULT_PAD_TOKEN
+from axolotl.utils.bench import log_gpu_memory_usage
 
 LOG = logging.getLogger("axolotl")
 
@@ -324,6 +325,9 @@ def load_model(
         )
         model.config.max_position_embeddings = cfg.sequence_len
 
+    if model.device.type == "cuda":
+        log_gpu_memory_usage(LOG, "after model load", model.device)
+
     if not cfg.gptq and (
         (cfg.adapter == "lora" and load_in_8bit)
         or (cfg.adapter == "qlora" and cfg.load_in_4bit)
@@ -360,6 +364,9 @@ def load_model(
                 module.scales = module.scales.half()
                 module.bias = module.bias.half()
 
+    if model.device.type == "cuda":
+        log_gpu_memory_usage(LOG, "after adapters", model.device)
+
     if (
         torch.cuda.device_count() > 1
         and int(os.getenv("WORLD_SIZE", "1")) > 1

src/axolotl/utils/trainer.py

@@ -18,6 +18,7 @@
 from transformers.trainer_pt_utils import get_parameter_names
 
 from axolotl.utils.callbacks import (
+    PrintGPUStatsCallback,
     SaveBetterTransformerModelCallback,
     SavePeftModelCallback,
 )
@@ -292,6 +293,7 @@ def setup_trainer(cfg, train_dataset, eval_dataset, model, tokenizer):
         trainer_kwargs["optimizers"] = (optimizer, lr_scheduler)
 
     callbacks = []
+    callbacks.append(PrintGPUStatsCallback(cfg))
     # TODO on_save callback to sync checkpoints to GCP/AWS in background
     if cfg.early_stopping_patience:
         early_stop_cb = EarlyStoppingCallback(