Hugging Face's logo

mlboydaisuke
/
qwen3.5-0.8B-CoreAI

Text Generation
coreai
coreai-aimodel
aimodel
apple-silicon
ane
on-device
qwen3.5
hybrid-ssm
gated-deltanet
Model card Files
xet
 Community
qwen3.5-0.8B-CoreAI

Commit History

30 days ago

  1. gen-cards: regenerate Use-it block

    mlboydaisuke
    Verified signature
    b98ad07

  1. gen-cards: regenerate Use-it block

    mlboydaisuke
    Verified signature
    639ba55

  1. Card: macOS/iOS 27 GA wording (beta requirement dropped; beta findings dated)

    mlboydaisuke
    Verified signature
    d368e55

  2. qwen3_5_0_8b_decode_int8.aimodel: re-save with coreai-core 1.0.0b2 (strip debug locations; weights unchanged). The 0.4.0-era IR refused AIModel.load on every OS 27 build since beta 2, still on the RC 26A428 (2026-09-14); the re-saved bundle loads. Recovery: coreai-model-zoo conversion/recovery/{strip_b1,resave_b2}.py

    mlboydaisuke
    Verified signature
    bdca16d

  3. qwen3_5_0_8b_ios_hc_prefill_q16_b2048_int8.aimodel: re-save with coreai-core 1.0.0b2 (strip debug locations; weights unchanged). The 0.4.0-era IR refused AIModel.load on every OS 27 build since beta 2, still on the RC 26A428 (2026-09-14); the re-saved bundle loads. Recovery: coreai-model-zoo conversion/recovery/{strip_b1,resave_b2}.py

    mlboydaisuke
    Verified signature
    7016bba

  4. qwen3_5_0_8b_ios_hc0_int8v3.aimodel: re-save with coreai-core 1.0.0b2 (strip debug locations; weights unchanged). The 0.4.0-era IR refused AIModel.load on every OS 27 build since beta 2, still on the RC 26A428 (2026-09-14); the re-saved bundle loads. Recovery: coreai-model-zoo conversion/recovery/{strip_b1,resave_b2}.py

    mlboydaisuke
    Verified signature
    753aeef

  5. qwen3_5_0_8b_ios_hc0.aimodel: re-save with coreai-core 1.0.0b2 (strip debug locations; weights unchanged). The 0.4.0-era IR refused AIModel.load on every OS 27 build since beta 2, still on the RC 26A428 (2026-09-14); the re-saved bundle loads. Recovery: coreai-model-zoo conversion/recovery/{strip_b1,resave_b2}.py

    mlboydaisuke
    Verified signature
    808639c

  6. qwen3_5_0_8b_decode_int8.aimodel: re-save with coreai-core 1.0.0b2 (strip debug locations; weights unchanged). The 0.4.0-era IR refused AIModel.load on every OS 27 build since beta 2, still on the RC 26A428 (2026-09-14); the re-saved bundle loads. Recovery: coreai-model-zoo conversion/recovery/{strip_b1,resave_b2}.py

    mlboydaisuke
    Verified signature
    51fcb11

  7. qwen3_5_0_8b_decode_int8lin: re-save with coreai-core 1.0.0b2 (strip debug locations; weights unchanged). The 0.4.0-era IR refused AIModel.load on every OS 27 build since beta 2, still on the RC 26A428 (2026-09-14); the re-saved bundle loads. Recovery: coreai-model-zoo conversion/recovery/{strip_b1,resave_b2}.py

    mlboydaisuke
    Verified signature
    ebb6509

  8. qwen3_5_0_8b_decode_int8hu_perchan_sym: re-save with coreai-core 1.0.0b2 (strip debug locations; weights unchanged). The 0.4.0-era IR refused AIModel.load on every OS 27 build since beta 2, still on the RC 26A428 (2026-09-14); the re-saved bundle loads. Recovery: coreai-model-zoo conversion/recovery/{strip_b1,resave_b2}.py

    mlboydaisuke
    Verified signature
    35d2e4f

  1. gen-cards: regenerate Use-it block

    mlboydaisuke
    Verified signature
    67acce1

  2. Card: DeviceMark row (2026-09)

    mlboydaisuke
    Verified signature
    d6b3c00

  1. Card metadata: library_name coreai, ensure coreai tag

    mlboydaisuke
    Verified signature
    f93560b

  1. Card: definition line (Core AI, 2026-09)

    mlboydaisuke
    Verified signature
    3cb436b

  1. Card: base_model_relation: quantized (so the conversion lists under the base model's Quantizations, not Finetunes)

    mlboydaisuke
    Verified signature
    df983f6

  1. gen-cards: regenerate Use-it block

    mlboydaisuke
    Verified signature
    342fa02

  1. Link the card back to its collection and the request box

    mlboydaisuke
    Verified signature
    e587505

  1. gen-cards: regenerate Use-it block

    mlboydaisuke
    Verified signature
    c50a10c

  1. Re-export with coreai-torch 0.4.1 (fixes OS 27 beta 2+ load failure); gated token-for-token vs fp32 oracle

    mlboydaisuke
    Verified signature
    1b8c020

  1. Note b2 (beta3-loadable) bundles under gpu-pipelined-b2/

    mlboydaisuke
    Verified signature
    0aa1653

  2. Add b2 (beta3-loadable) bundle under gpu-pipelined-b2/qwen3_5_0_8b_decode_int8hu_block32_sym/ (b1 retained)

    mlboydaisuke
    Verified signature
    3c154ef

  1. gen-cards: regenerate Use-it block

    mlboydaisuke
    Verified signature
    34ed8b0

  2. gen-cards: add Use-it markers

    mlboydaisuke
    Verified signature
    4c804ba

  1. Add root config.json (Hub download-stats query file + framework pointer)

    mlboydaisuke
    Verified signature
    81e1928

  1. head-quant docs: per-block-32 absmax ship shape (per-channel = beta delegate bug; naming note)

    mlboydaisuke
    Verified signature
    4447ec6

  2. Upload folder using huggingface_hub

    mlboydaisuke
    Verified signature
    95388e0

  3. Upload README.md with huggingface_hub

    mlboydaisuke
    Verified signature
    d97bd51

  1. card: gpu-pipelined int8lin bundle (iPhone 50.3-51.5 / Mac 204 tok/s) + run instructions

    mlboydaisuke
    Verified signature
    b647bda

  2. qwen3.5-0.8B int8lin decode-only loop-free bundle (pipelined engine): Mac 204 tok/s, iPhone 50.3-51.5 tok/s

    mlboydaisuke
    Verified signature
    0019cb6

  3. int8 fused-kernel monolith (42.5-45.4 tok/s) + q16 chunked-prefill companion (147 tok/s) — new release config

    mlboydaisuke
    Verified signature
    c2a8a57

  4. ios-gpu: add qwen3_5_0_8b_ios_hc_prefill_q16_b2048_int8.aimodel

    mlboydaisuke
    Verified signature
    23e3da6

  5. ios-gpu: add qwen3_5_0_8b_ios_hc0_int8v3.aimodel

    mlboydaisuke
    Verified signature
    b0080fd

  6. Remove pre-category-layout path ios-gpu-static/

    mlboydaisuke
    Verified signature
    161130e

  7. Remove pre-category-layout path dynamic-int8/

    mlboydaisuke
    Verified signature
    7eb65df

  8. macOS GPU best: dynamic int8 (58.5 tok/s release)

    mlboydaisuke
    Verified signature
    cc382cb

  9. iOS ANE best: dynamic int8 (14.7 tok/s)

    mlboydaisuke
    Verified signature
    f35903b

  10. iOS GPU best: fp16 static ctx-2048 monolith (27.7 tok/s)

    mlboydaisuke
    Verified signature
    2d7f93a

  11. Card: category layout (best verified config per platform x compute-unit)

    mlboydaisuke
    Verified signature
    7f8783c

  12. static ctx-2048 monolith (iPhone GPU 27.7 tok/s, release config)

    mlboydaisuke
    Verified signature
    a94f01c

  13. dynamic int8 bundle (iPhone GPU 12.5 / ANE 14.7 / Mac 58.5 tok/s)

    mlboydaisuke
    Verified signature
    48e0a23

  14. Model card

    mlboydaisuke
    Verified signature
    7ae2c03

  15. initial commit

    mlboydaisuke
    Verified signature
    513c05f