304b8eca7a* fix: match qwen3-thinking chat template double-newline in response pattern The Qwen3-thinking chat template generates `<think>\n\n` (double newline) after the think tag, but `train_on_responses_only` was looking for `<think>\n` (single newline). `\n\n` is token 271 while `\n` is token 198 -- different tokens, so the pattern match in `train_on_responses_only` fails, masking ALL tokens and dropping 100% of training samples. Update the response pattern from `<think>\n` to `<think>\n\n` to match what the actual qwen3-thinking template generates. Fixes #6919 * fix qwen3 thinking response marker --------- Co-authored-by: Ayushman Paul <ayushman@HP> Co-authored-by: Etherll <61019402+Etherll@users.noreply.github.com> |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| cache_safe.py | ||
| chat_templates.py | ||
| data_collators.py | ||
| dataset_none_detect.py | ||
| dataset_utils.py | ||
| format_conversion.py | ||
| format_detection.py | ||
| iterable.py | ||
| llm_assist.py | ||
| model_mappings.py | ||
| raw_text.py | ||
| vlm_processing.py | ||