GitLab

Update DeepseekR1_V3_tutorial.md

Support chunk prefill. Support 139K context for DeepSeek-R1 139K with in 24G VRAM.

fix cache_lens bug in server and rm test prompt.txt

Delete duplicate code

⚡ fox docker build

[fix] Fix template name

[UPDATE] Update ZH/EN issue template

fix temperature

use generation config from json file in official repo

Allow temperature and top_p from /v1/chat/completions

Update issue templates

Fix missing macro definition for KTRANSFORMERS_USE_CUDA and <chrono> includes on MSVC

Fix RuntimeError on Windows caused by integer overflow in np.prod

fix: fix SSE formatting

feat: basic api key support

Feat: Clear cache during weight loading to prevent OOM on GPUs with <=8GB VRAM

feat:implementation of chat routing for Ollama

[UPDATE] Update documents.

Update DeepseekR1_V3_tutorial.md

Update DeepSeek-V3-Chat-multi-gpu-marlin.yaml