SenseNova-Loop-9B-NVFP4 / hf_quant_config.json

Commit History

Yesterday

  1. Quantize the LM head to NVFP4 (22 % faster decode, 2 GB smaller), add MTP and AWQ comparison to the card

    Verified signature
    659937b

2 days ago

  1. Add hf_quant_config.json

    Verified signature
    372ef1b