GitLab

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

fix numa cpu distribution

fix dockerfile in devcontainer and fix expert torch

The numa node location would be calculated based on the total number
of worker threads.
So we should always use the actual number of threads instead of using a min() op.