Commit History

27 days ago

  1. Add tested Qwen3.5 2B XNNPACK and MLX chat exports⁠

    Verified signature
    73b03b8

  1. docs: update model card, context lengths, and config specs

    b94c758

  1. Stop restating quantized, and drop the unused default⁠

    0fe1c0b

  1. Update chat template for strict monotonicity and tool support

    7f4a202

  1. Use min/max dynamic shape bounds for forward inputs/outputs

    8ef2f7f
  2. Backfill constant metadata method values in config.json from .pte

    d9dff69

  1. fix(v0.9.0): split text vs VL into <size>/ and vl_<size>/; colocate tokenizers per-size; drop Qwen3.5-* legacy dirs

    Verified signature
    f296584

  1. Add spec-compliant config.json files

    Verified signature
    1e50a8e
  2. Add spec-compliant config.json files

    Verified signature
    48c2dab
  3. Restructure to MODEL_SPEC.md convention

    Verified signature
    0ee1d48