Issues · Blaizzy/mlx-vlm · GitHub

Batch Processing Feature
#40 · Blaizzy opened on Jun 11, 2024
8
ChatUI improvements
#45 · Blaizzy opened on Jun 23, 2024
Models to port to MLX-VLM
#39 · Blaizzy opened on Jun 11, 2024
58

Labels Milestones New issue

mlx_vlm.convert with mtp?

#1222

· chigkim opened

on May 26, 2026

Restore TurboQuant fused K/V decode path for RHT codecs

#1220

· Blaizzy opened

on May 25, 2026

TypeError: _build_replacement_call got an unexpected keyword argument 'target_verify' in Qwen3.5/3.6 MTP models — PR #1210 does not resolve

#1219

· cropduster opened

on May 25, 2026

20260504 -- Diagnostics Report — 3 failure(s), 9 harness issue(s) (mlx-vlm 0.5.0)

#1218

· jrp2014 opened

on May 24, 2026

20260515 -- Diagnostics Report — 3 failure(s), 13 harness issue(s) (mlx-vlm 0.5.0+)

#1187

· jrp2014 opened

on May 15, 2026

MLX APC Cache Not Hitting for Specific Prompt (Exact Hits Always 0)

#1185

· NeoInBJ opened

on May 15, 2026

Support APC with KV-cache quantization (critical for TTFT & long Agent tasks for local-first AI)

#1174

· Edsuns opened

on May 12, 2026

Support response_format type=json_object

#1171

· eloe opened

on May 12, 2026

20260508 -- Diagnostics Report — 7 failure(s), 2 harness issue(s) (mlx-vlm 0.5.0)

#1156

· jrp2014 opened

on May 8, 2026

Add tools / function-calling support to /v1/responses

#1136

· jasonvassallo opened

on May 7, 2026

No model is included in the JSON response when making a request to `http://127.0.0.1:8080/v1/models`

#1133

· slushhh opened

on May 6, 2026

mlx_vlm.server crashes if the `~/.cache/huggingface/hub` directory is missing

#1132

· slushhh opened

on May 6, 2026