v0.1.4

v0.1.4 Release

Bug Fixes

• Fix chat template warmup issue (#253)

Features

• Support SGLang v0.5.9
• Support vLLM v0.16.0 (#243)
• Support Python 3.13 (#260)
• Support MLA models with multi-page-size (#240)
• Support ElasticMLAMemoryPool (#238)
• Support SGLang page-size>1 (#237)
• Support GPT-OSS in SGLang-v0.5.6 (#249)

Maintenance

• Delete redundant functions in SGLang patches (#239)
• Update the launch script (#255)
• Update ignorance of vscode settings (#244)

Documentation

• Add Updates section in README (#251)
• Update launch script

v0.1.3

v0.1.3 Release

Bug Fixes

• Fix unique IPC socket when TP is enabled (#219)
• Support flashinfer kvcache shape for vLLM (#216)
• Fix exceptions in vLLM prefix-cache-related APIs (#212)
• Export missing autopatch env (#196)
• More precise vLLM version detection (#194)
• Fix memory tracker IPC name collision (#192)
• Fix type casting in get_avail_physical_pages (#190)
• Fix vLLM version range (#182)
• Fix SGLang clear() function support (#150)
• Fix langchain AgentExecutor import (#209)
• Fix type casting issue (#204)

Features

• Add example for hybrid attention models (#218)

Documentation

• Add deepwiki to README (#220)
• Update README with kvcached caching note (#188)
• Add action video and blog link (#178)

Maintenance

• kvcached description update (#180)
• Example READMEs venv path fix (#179)

v0.1.2

• Fix KV cache allocation bugs
• Enhance examples
• Add GSMK8 benchmark
• Update docs and dockers

v0.1.1

The first official release of kvcached!