423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-08-02 19:52:50 +08:00
Code Issues Packages Projects Releases Wiki Activity

[trainer] fix KeyError at end of pretrain (#8099)

Browse Source
This commit is contained in:
Ma, Xiaochen 2025-05-19 18:01:26 +08:00 committed by GitHub
parent 89a0d10c18
commit 16e26236eb
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 15 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/llamafactory/train/pt/workflow.py
View File

@ -77,12 +77,22 @@ def run_pt(
# Evaluation # Evaluation
if training_args.do_eval: if training_args.do_eval:
metrics = trainer.evaluate(metric_key_prefix="eval") metrics = trainer.evaluate(metric_key_prefix="eval")
if isinstance(dataset_module.get("eval_dataset"), dict):
for key in dataset_module["eval_dataset"].keys():
try:
perplexity = math.exp(metrics[f"eval_{key}_loss"])
except OverflowError:
perplexity = float("inf")
metrics[f"eval_{key}_perplexity"] = perplexity
else:
try: try:
perplexity = math.exp(metrics["eval_loss"]) perplexity = math.exp(metrics["eval_loss"])
except OverflowError: except OverflowError:
perplexity = float("inf") perplexity = float("inf")
metrics["eval_perplexity"] = perplexity
metrics["perplexity"] = perplexity
trainer.log_metrics("eval", metrics) trainer.log_metrics("eval", metrics)
trainer.save_metrics("eval", metrics) trainer.save_metrics("eval", metrics)