mlx-examples · Issues· 180 open

Contribute Hugging Face models to the MLX Community

segment_anything: fuse attention with mx.fast.scaled_dot_product_attention

Community project: Depth-Anything-V2 running via a from-scratch PyTorch-API shim (torch-mlx)

Proposal: neural surface-code decoder (QEC) example?

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

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

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

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

mlx_lm.fuse can not work

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

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

whisper: align no_speech_threshold fallback condition with openai/whisper

[Feature Request] LLaVA 1.6 LoRA fine-tuning example

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

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