freeliuzc d49f8fb30a [Feature][MTP] Support cacheKV transfer in per_chunk mode (#2890) ...

* support chunk_prefill both normal and speculative_decoding(mtp)

* optimize pd-disaggregation config

* fix bug

2025-07-17 17:58:08 +08:00

..

attention

[Feature][MTP] Support cacheKV transfer in per_chunk mode (#2890)

2025-07-17 17:58:08 +08:00

backends

refactor rl get_name_mappings_to_training (#2847)

2025-07-15 07:31:42 -07:00

moe

rl update (#2861)

2025-07-16 00:33:10 -07:00

quantization

refactor rl get_name_mappings_to_training (#2847)

2025-07-15 07:31:42 -07:00

sample

[Feature] support prompt repetition_penalty (#2806)

2025-07-17 12:05:52 +08:00

__init__.py

[LLM] First commit the llm deployment code

2025-06-09 19:20:15 +08:00

activation.py

[GCU] Support gcu platform (#2702)

2025-07-08 13:00:52 +08:00

embeddings.py

Fix rollout_model init (#2881)

2025-07-16 22:36:21 -07:00

linear.py

refactor rl get_name_mappings_to_training (#2847)

2025-07-15 07:31:42 -07:00

lm_head.py

refactor rl get_name_mappings_to_training (#2847)

2025-07-15 07:31:42 -07:00

mtp_linear.py

refactor rl get_name_mappings_to_training (#2847)

2025-07-15 07:31:42 -07:00

normalization.py

refactor rl get_name_mappings_to_training (#2847)

2025-07-15 07:31:42 -07:00

rotary_embedding.py

fix and refine vl (#2866)

2025-07-16 05:59:28 -07:00

utils.py

Adapt for iluvatar gpu (#2684)

2025-07-07 16:53:14 +08:00