Sign up

Shiftedx/Qwythos-27B-v1-MXFP8-Vision-MLX

🤗 Hugging Face sourceimage-text-to-textapache-2.027.4B params55 GBsafetensors✓ Checksums witnessedupdated today
No torrent yet →

Qwythos-27B-v1-MXFP8-Vision-MLX

MLX/VLM conversion of empero-ai/Qwythos-27B-v1, pinned to revision 7c72a9c714cf66281cb222c4aa0aef368d84c94f.

Format

  • Language body: MXFP8, 8-bit, group size 32
  • Vision tower: source BF16, 333 tensors
  • Architecture: dense Qwen3.5 multimodal, 64 language layers
  • Chat template and tokenizer: preserved from the pinned source

This standard artifact intentionally disables the embedded MTP layer and does not include mtp.safetensors.

Local qualification

  • Structural inspection: pass (1,682 indexed tensors; 498 quantized weights; no stale or missing shards)
  • Strict MLX-VLM load: pass
  • Text behavior: 5/5 exact matches against BF16 for identity, JSON, code, tool calling, and loop tasks
  • Vision smoke: pass; post-graft vision features are bit-for-bit equal to BF16
  • Long-context needle: exact at 19,069 input tokens

The source configuration's 1M-token YaRN settings are preserved, but this release was locally exercised only through 19,069 input tokens—not at 1M.

Basic use

python -m mlx_vlm.generate \
  --model <local-path-or-repo-id> \
  --image image.jpg \
  --prompt "Describe this image." \
  --max-tokens 256

Review the source model card for intended use, training, license, and deployment considerations. This conversion does not change the source model's safety characteristics.