Commit History

  1. docs: align model card to 0.3.1 (honest wins-only: 2.6x routing, native-default, English-only, no GPU/Windows-native/end-to-end claims)

    Verified signature
    016e5a4

  1. docs: align model card with v0.3.0b9 (hook, acquire, native encoder, pure-MIND migration)

    Verified signature
    27e6629

  1. chore: yank encoder_weights.q16.meta.json — superseded; v0.3.0b7 had a silent correctness regression (i64/i32 stride mismatch in dot_q16_v); v0.3.0b8 hotfixes by reverting to Track A C-shim

    Verified signature
    47fa73e
  2. chore: yank encoder_weights.q16.bin — superseded; v0.3.0b7 had a silent correctness regression (i64/i32 stride mismatch in dot_q16_v); v0.3.0b8 hotfixes by reverting to Track A C-shim

    Verified signature
    2db2558
  3. feat: encoder_weights meta sidecar (mirrors the (N,K) blob)

    Verified signature
    ae5da41
  4. feat: encoder_weights.q16.bin (N,K) layout for mind-nerve 0.3.0b7 #233(a) thesis-pure encode (sha 17489317...)

    Verified signature
    8183455

  1. docs: highlight pip-install + auto-download (0.1.0a6)

    Verified signature
    71221fd
  2. mind-nerve phase1 v1.1-oss — initial weights upload (0.1.0a3)

    Verified signature
    bc96d99
  3. Upload README.md with huggingface_hub

    Verified signature
    318b43d
  4. initial commit

    Verified signature
    464a398