Ayakouji 453487d5b0 [Feat] ernie4_5_vl_moe support CudaGraph (#3226) ...

* delete dynamic control flow for decode

* coda-style

* fix scatter/gather typos and use input stream instead default stream

* support 0-Size Tensor

* update runner and model

* using static mem address as input

* fix mem leak

* refine code

* update mm_buffer

* fix typo

* fix buffersize

* fix unk token

* refine code

* refine

* support other arch

* open cudagraph in vlci

* fix

* update

* update

* update

* fix cmd

* update

---------

Co-authored-by: aquagull <hongyuh@qq.com>
Co-authored-by: Yuanle Liu <yuanlehome@163.com>

2025-09-10 13:11:57 +08:00

..

graph_optimization

[Feat] ernie4_5_vl_moe support CudaGraph (#3226)

2025-09-10 13:11:57 +08:00

guided_decoding

[Feature] mm and thinking model support structred output (#2749)

2025-09-02 16:21:09 +08:00

layers

[BugFix] Fix the abnormal memory usage caused by shape errors in the triton moe backend (#4026)

2025-09-09 20:05:54 -07:00

model_loader

cache feature (#3857)

2025-09-07 18:52:46 +08:00

models

[Feat] ernie4_5_vl_moe support CudaGraph (#3226)

2025-09-10 13:11:57 +08:00

ops

fix cpu __ini__.py (#3448)

2025-08-17 12:38:54 +08:00

__init__.py

polish code with new pre-commit rule (#2923)

2025-07-19 23:19:27 +08:00

forward_meta.py

[Excutor] Experiment Feature-Support Prefill in cudagraph (#3459)

2025-09-08 13:12:24 +08:00

load_weight_utils.py

cache feature (#3857)

2025-09-07 18:52:46 +08:00

pre_and_post_process.py

support mtp in v1_scheduler mode (#3695)

2025-09-04 17:39:59 +08:00

utils.py

cache feature (#3857)

2025-09-07 18:52:46 +08:00