[HybridKV] Fix prefill disaggregation kvcache addr alignment & use hybrid kv cache only when running qwen3_next #3981
image_a3_ubuntu.yml
on: pull_request
vllm-ascend image build
16m 16s
Artifacts
Produced during runtime
Name | Size | Digest | |
---|---|---|---|
vllm-project~vllm-ascend~4XQFDB.dockerbuild
|
106 KB |
sha256:4bbd2f8f4aea36d9452e76cb6d5665a9253a05f4736a6ce268dfb5f531087e1d
|
|