COMMITS
March 19, 2026
W
Refresh requirements.txt (#292)
will-lms committed
February 20, 2026
N
Add max tokens test (#281)
Neil Mehta committed
February 19, 2026
N
[batched gen] Stop iterating when max tokens is reached (#279)
Neil Mehta committed
W
Set max_seq_nums to 1 in demo.py (#278)
will-lms committed
February 11, 2026
M
Add another image to test data that triggers failing test (#274)
Matt Clayton committed
February 10, 2026
M
Rebase mlx-vlm (#273)
Matt Clayton committed
February 9, 2026
N
Use ModelKit if max_seq_nums is explicitly set to 1 (#272)
Neil Mehta committed
February 5, 2026
N
Add unit test for batched gen (#267)
Neil Mehta committed
N
Raise error when parallel batching for vision is configured (#268)
Neil Mehta committed
N
Add continuous batching support for text-only models (#266)
Neil Mehta committed
January 15, 2026
W
Fix heavy test (#260)
will-lms committed
W
Prompt Processing V2 Callback (#259)
will-lms committed
January 8, 2026
W
LFM 2.5 VL (#255)
will-lms committed
December 17, 2025
M
Fix LFM2-VL with new mlx-vlm (#253)
Matt Clayton committed
M
Force correct processor config for LM Studio Mistral Small 3.2 (#252)
Matt Clayton committed
December 16, 2025
M
Revert "Force correct processor config for LM Studio Mistral Small 3.2"
Matt Clayton committed
M
Force correct processor config for LM Studio Mistral Small 3.2
Matt Clayton committed
December 15, 2025
N
Fix mistral tokenization for transformers v5 rc1 (#249)
Neil Mehta committed
December 12, 2025
W
Broaden `eos_token_id` check (#250)
will-lms committed
December 11, 2025
N
Devstral Small 2 support; Transformers v5 upgrade (#248)
Neil Mehta committed
W
RNJ-1 EOS Correction (#247)
will-lms committed
November 13, 2025
N
Update load_tokenizer import (#242)
Neil Mehta committed
November 7, 2025
N
Add max-img-size arg to demo.py (#240)
Neil Mehta committed
October 27, 2025
N
Disable qwen 2 VL unified arch (#239)
Neil Mehta committed
October 23, 2025
N
disable qwen3-vl vision add ons (#238)
Neil Mehta committed
N
Log image resize only when it happens (#236)
Neil Mehta committed
N
Remove hard-coded image resize (#235)
Neil Mehta committed
October 15, 2025
V
Add Qwen3-VL (Dense) vision add-on (#234)
Vincent Amato committed
October 13, 2025
V
Add Qwen3-VL (MoE) vision add-on (#230)
Vincent Amato committed
N
Upgrade MLX requirements (#231)
Neil Mehta committed