Repository Issue Activity (beta)

#1443 Suggestion: add a Stable Fast 3D (SF3D) example — native MLX single-image-to-3D-mesh

#1441 [whisper] `mlx_whisper` CLI overwrites output file when multiple audio inputs are provided

#1440 whisper: numerical parity tests fail with MLX 0.32.0 on Apple Silicon

#1421 whisper/convert.py outputs `model.safetensors` but `mlx_whisper.load_models` expects `weights.safetensors`

#1436 mlx_lm.fuse can not work

#1435 Stable Diffusion: Transformer2D's GroupNorm uses MLX's default eps (1e-5) where diffusers uses 1e-6

#1434 Stable Diffusion: Transformer2D's GroupNorm uses MLX's default eps (1e-5) where diffusers uses 1e-6

#1433 [Proposal] DeepSeek-OCR example — SAM + CLIP dual vision encoder → DeepSeek-V2 MoE

#1427 whisper: align no_speech_threshold fallback condition with openai/whisper

#605 [Feature Request] LLaVA 1.6 LoRA fine-tuning example

#1418 [whisper] `KeyError: 'words'` on transcription (with `--task translate`)

#1417 Support `--prompt-cache-file` with mlx_lm.server

#1402 Question: (mlx_whisper) How to eliminate spurious "you" transcriptions

#1254 mlx_whisper memory usage keeps growing for every new audio chunk processed with word_timestamps on.

#1412 [whisper] Add Flash Attention and batched decoding for up to 10x speedup

#1373 Is there a way to implement a Timeout / cancellation for transcribe() method, espeicially model.decode()? (large-v3 sometimes stalls >200s on ~10s audio)

#1410 [whisper] Feature Request: Add `carry_initial_prompt` parameter to `transcribe()`

#866 M2 Ultra 192 GB fails to run while M3 Max 128GB can run

#1408 mlx_lm: TokenizerWrapper drops 'eos_token_ids' array, breaking conversational stop tokens

#1198 Concurrent requests

#1405 [Model Request] Porting Nvidia PersonaPlex-7B (Moshi architecture) & Mimi Codec

#664 libc++abi: terminating due to uncaught exception of type std::runtime_error

#1379 Support for Qwen-Image generation

#1397 N

#1029 Slow Speed Issue