6 projects
vllm-ascend-quantized-kv-cache
Runtime-loadable INT8 KV-cache attention backend for vLLM-Ascend-HUST
vllm-ascend-quant-ext
Fail-closed runtime support for versioned Ascend quantized model artifacts
vllm-ascend-kvcompress-hust
TriAttention KV-cache compression plugin for vLLM Ascend
vllm-hust-bidkv
BidKV — CompressionBid protocol layer for KV cache scheduling primitives
vllm-hust-vspec
Out-of-tree speculative decoding plugin for vLLM-HUST on Ascend
vllm-hust-engine
Complete frozen vLLM-HUST R9 runtime and model adapters