github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-09 15:08:02 -05:00
Code Issues Packages Projects Releases Wiki Activity

fix bert free_intermediates (#9633)

Browse Source 
fix when only run eval `TRAIN=0 BERT_SIZE=tiny examples/mlperf/training_submission_v5.0/tinycorp/benchmarks/bert/implementations/tinybox_green/dev_beam.sh`
This commit is contained in:
chenyu
2025-03-30 22:42:52 -04:00
committed by GitHub
parent ff984c807d
commit d8d7ac1bb1
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
examples/mlperf/model_train.py
View File

@@ -826,7 +826,7 @@ def train_bert():
if MLLOGGER and RUNMLPERF:
MLLOGGER.start(key=mllog_constants.EVAL_START, value=None, metadata={"epoch_num": i*BS, "step_num": i})
if getenv("RESET_STEP", 0): train_step_bert.reset()
else: train_step_bert.captured.free_intermediates()
elif train_step_bert.captured is not None: train_step_bert.captured.free_intermediates()
eval_lm_losses = []
eval_clsf_losses = []
eval_lm_accs = []
@@ -860,7 +860,7 @@ def train_bert():
return
if getenv("RESET_STEP", 0): eval_step_bert.reset()
else: eval_step_bert.captured.free_intermediates()
elif eval_step_bert.captured is not None: eval_step_bert.captured.free_intermediates()
del eval_data
avg_lm_loss = sum(eval_lm_losses) / len(eval_lm_losses)
avg_clsf_loss = sum(eval_clsf_losses) / len(eval_clsf_losses)