ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix: comment out debug print statements (#4357)

Browse source
This commit is contained in:
Roland Tannous 2026-03-17 15:43:27 +04:00 • committed by GitHub
commit a0aba96ebd
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 34 additions and 34 deletions
Download patch file Download diff file Expand all files Collapse all files

38
studio/backend/core/training/trainer.py
View file

@ -3270,25 +3270,25 @@ class UnslothTrainer:
logger.info(f"Post-filter dataset size: {filtered_len} samples\n")
# [DEBUG] Decode first sample AFTER train_on_completions applied
try:
_row = self.trainer.train_dataset[0]
_space = self.tokenizer(
" ", add_special_tokens = False
).input_ids[0]
print("[DEBUG] === After train_on_completions ===", flush = True)
print(
f"[DEBUG] input_ids decoded:\n{self.tokenizer.decode(_row['input_ids'])}\n",
flush = True,
)
print(
f"[DEBUG] labels decoded (-100 → space):\n{self.tokenizer.decode([_space if x == -100 else x for x in _row['labels']])}\n",
flush = True,
)
except Exception as _dbg_e:
print(
f"[DEBUG] Could not decode post-completions sample: {_dbg_e}",
flush = True,
)
# try:
# _row = self.trainer.train_dataset[0]
# _space = self.tokenizer(
# " ", add_special_tokens = False
# ).input_ids[0]
# print("[DEBUG] === After train_on_completions ===", flush = True)
# print(
# f"[DEBUG] input_ids decoded:\n{self.tokenizer.decode(_row['input_ids'])}\n",
# flush = True,
# )
# print(
# f"[DEBUG] labels decoded (-100 → space):\n{self.tokenizer.decode([_space if x == -100 else x for x in _row['labels']])}\n",
# flush = True,
# )
# except Exception as _dbg_e:
# print(
# f"[DEBUG] Could not decode post-completions sample: {_dbg_e}",
# flush = True,
# )
except Exception as e:
logger.warning(f"Failed to apply train on responses only: {e}")

30
studio/backend/core/training/worker.py
View file

@ -314,21 +314,21 @@ def run_training_process(
# [DEBUG] Print first sample before model is loaded
# dataset is a dict {"dataset": <Dataset>, "detected_format": ..., ...}
# or a raw Dataset for audio paths
try:
ds = dataset["dataset"] if isinstance(dataset, dict) else dataset
print(
f"\n[DEBUG] Dataset loaded BEFORE model. type={type(ds).__name__}, len={len(ds)}",
flush = True,
)
print(f"[DEBUG] Columns: {ds.column_names}", flush = True)
sample = ds[0]
preview = {k: str(v)[:300] for k, v in sample.items()}
print(f"[DEBUG] First sample: {preview}\n", flush = True)
except Exception as e:
print(
f"[DEBUG] Could not preview first sample: {type(e).__name__}: {e}",
flush = True,
)
# try:
# ds = dataset["dataset"] if isinstance(dataset, dict) else dataset
# print(
# f"\n[DEBUG] Dataset loaded BEFORE model. type={type(ds).__name__}, len={len(ds)}",
# flush = True,
# )
# print(f"[DEBUG] Columns: {ds.column_names}", flush = True)
# sample = ds[0]
# preview = {k: str(v)[:300] for k, v in sample.items()}
# print(f"[DEBUG] First sample: {preview}\n", flush = True)
# except Exception as e:
# print(
# f"[DEBUG] Could not preview first sample: {type(e).__name__}: {e}",
# flush = True,
# )
# Disable eval if eval_steps <= 0
eval_steps = config.get("eval_steps", 0.00)