Skip to content

Actions: janhilgard/vllm-mlx

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
27 workflow runs
27 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
feat: live inference KV cache quantization with asymmetric K/V bits CI #27: Commit 9dce02c pushed by janhilgard
2m 28s main
main
2m 28s
split summarize_workload_results into per-case helper (#519) CI #26: Commit 0140ae0 pushed by janhilgard
2m 23s main
main
2m 23s
fix: add missing _eval_prompt_cache and PrefillAbortedError definitions CI #25: Commit 011a0af pushed by janhilgard
1m 28s main
main
1m 28s
feat: add reasoning parser support to Anthropic Messages API CI #24: Commit e5d0052 pushed by janhilgard
1m 17s main
main
1m 17s
fix: run MLLM prefill in thread executor to keep event loop responsive CI #23: Commit 8b9bc86 pushed by janhilgard
1m 22s main
main
1m 22s
Fix RotatingKVCache merge crash on long prompts (>sliding_window) CI #22: Commit 2d2d516 pushed by janhilgard
1m 4s main
main
1m 4s
Add per-request status info for MLLM scheduler CI #21: Commit ccee11b pushed by janhilgard
1m 8s main
main
1m 8s
style: run black on test_gemma4_parser.py and qwen3_5_mtp.py CI #20: Commit c2d6b04 pushed by janhilgard
1m 25s main
main
1m 25s
fix: remove unused imports and extraneous f-string prefixes (lint) CI #19: Commit 06fb183 pushed by janhilgard
1m 2s main
main
1m 2s
Add Gemma 4 tool call parser (from PR #254) CI #18: Commit b94fbae pushed by janhilgard
1m 21s main
main
1m 21s
style: Run black on simple.py CI #17: Commit 4d2b026 pushed by janhilgard
1m 3s main
main
1m 3s
fix: Update Qwen3 parser tests to match fork reasoning behavior CI #16: Commit e12a1b7 pushed by janhilgard
59s main
main
59s
fix: Remove duplicate --mllm argparse definition CI #15: Commit 6b543d1 pushed by janhilgard
1m 13s main
main
1m 13s
Merge remote-tracking branch 'origin/main' into merge/upstream-to-fork CI #14: Commit 1d0907c pushed by janhilgard
1m 16s main
main
1m 16s
fix: replace hard prefill token limit with warning CI #13: Commit fc34b87 pushed by janhilgard
1m 1s main
main
1m 1s
fix: enable_thinking for reasoning models + MLLM stats promotion CI #12: Commit 8b5cfe7 pushed by janhilgard
1m 18s main
main
1m 18s
fix: MLLM image processing — exclude_none for Jinja template, error h… CI #11: Commit 310246d pushed by janhilgard
1m 5s main
main
1m 5s
feat: patch Qwen3.5 attention for BatchKVCache compatibility CI #10: Commit bf78650 pushed by janhilgard
1m 20s main
main
1m 20s
feat: Support hybrid cache (ArraysCache+KVCache) in MLLM batching CI #9: Commit 8eb783f pushed by janhilgard
1m 30s main
main
1m 30s
style: black formatting CI #8: Commit d27c5ce pushed by janhilgard
1m 19s main
main
1m 19s
fix: Remove unused import and f-string prefix (lint) CI #7: Commit cfcf32a pushed by janhilgard
1m 16s main
main
1m 16s
feat: Qwen3.5 VLM loading, streaming detokenizer, tool markup stripping CI #6: Commit 476f717 pushed by janhilgard
1m 21s main
main
1m 21s
Merge upstream main (PR #53: GPT-OSS reasoning parser) CI #5: Commit ae35736 pushed by janhilgard
1m 23s main
main
1m 23s
Merge branch 'main' of https://github.com/janhilgard/vllm-mlx CI #4: Commit cac39be pushed by janhilgard
3m 54s main
main
3m 54s
fix: Clear stale MTP state between requests to prevent generation hang CI #3: Commit a18684b pushed by janhilgard
1m 7s main
main
1m 7s