Hugging Face's logo

majentik
/
Leanstral-RotorQuant-MLX-8bit

Text Generation
MLX
Safetensors
mistral3
rotorquant
kv-cache-quantization
8bit
weight-quantization
leanstral
lean4
formal-proofs
theorem-proving
quantized
apple-silicon
mistral
Mixture of Experts
8-bit precision
Model card Files
xet
 Community
Leanstral-RotorQuant-MLX-8bit

Commit History

  1. Remove links to retired duplicate brand variants

    majentik
    Verified signature
    8d74100

  1. Card accuracy pass 3: remove brand-comparison claim tables

    majentik
    Verified signature
    391ba0d

  2. Card accuracy pass 2: remove unmeasured speed claims, honest brand labels

    majentik
    Verified signature
    d1dec42

  3. Card accuracy sweep: honest brand labeling, remove dead links, upstream KV tip

    majentik
    Verified signature
    ffc8ab5

  4. Remove links to retired card-only KV-docs repos

    majentik
    Verified signature
    4e0de35

  1. docs: upstream-first KV-cache guidance (q8_0/q4_0, mainline Hadamard rotation); fork demoted to experimental

    majentik
    Verified signature
    76d45b9

  1. docs: Tier 2 polish — variant matrix + quant trade-off

    majentik
    Verified signature
    4811291

  1. chore(card): enrich YAML frontmatter (pipeline_tag, language, library_name, inference)

    majentik
    Verified signature
    fa36152

  1. Add MLX 8-bit quantized model with KV cache compression

    majentik
    Verified signature
    cfb84ae

  2. initial commit

    majentik
    Verified signature
    1f0f7ce