Chranos ebdc6fed03 fix: pass lm_head to LogitsProcessor instead of calling forward() ...

In vLLM v0.6.2, ParallelLMHead.forward() raises RuntimeError since
its weights should be used through LogitsProcessor.linear_method.apply().
Pass lm_head as first arg to LogitsProcessor which handles the
hidden_states -> logits projection internally.

2026-02-06 14:21:14 +08:00

..

__init__.py

testing dynamic register

2026-02-05 18:02:59 +08:00

base.py

testing dynamic register

2026-02-06 14:17:06 +08:00

causal.py

fix: pass lm_head to LogitsProcessor instead of calling forward()

2026-02-06 14:21:14 +08:00

legacy.py

testing dynamic register

2026-02-05 18:02:59 +08:00

pooling.py

testing dynamic register

2026-02-05 18:02:59 +08:00

utils.py

testing dynamic register

2026-02-06 14:17:06 +08:00