4 people 1b2b0a06e1 fix(learning): exclude padding tokens from SFT loss (#521) ...

* fix(learning): exclude padding tokens from SFT loss

* test(learning): add regression tests for SFT padding-loss masking

Cover the fix in both trainers (#521):
- OrchestratorSFTDataset.__getitem__: labels are -100 at padded positions,
  equal to input_ids elsewhere, and input_ids is not mutated.
- LoRATrainer._train_step: the labels passed to the model are masked at
  padded positions (captured via an injected model), with input_ids intact.

Both are torch-gated (pytest.importorskip / skipif HAS_TORCH), matching the
project's existing torch test gating, so they skip cleanly in the default CI
env. Verified locally with CPU torch: both PASS against the fix and FAIL
against the pre-fix code.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Jon Saad-Falcon <jonsaadfalcon@gmail.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

2026-06-10 11:48:57 -07:00

..

__init__.py

init commit

2026-03-12 17:29:39 +00:00

test_data.py

fix(channels): wire channel→agent handler and fix Telegram send pipeline (#94)

2026-03-20 18:35:18 -07:00

test_lora.py

fix(learning): exclude padding tokens from SFT loss (#521)

2026-06-10 11:48:57 -07:00