Hugging Face's logo

Spaces:
richardlin823
/
modules_play
Running

App Files Community
modules_play

Commit History

  1. feat: update skills.

    richard.lin
    2501836

  2. feat: add unidrive_vla_trt_edge_llm fp16 test, but failed

    richard.lin
    b26166d

  3. fix: update tencent dev machine's ip

    richard.lin
    4b25e3b

  4. feature: claude skills

    richard.lin
    766bec2

  5. fix: onnx export, consolidate the model onto one device.

    richard.lin
    4fa86e6

  1. fix: unidrive vla quantization, oom failure during calib. add dynamic loading of calibration data, add max_pixel to limit mass loading.

    richard.lin
    945cba1

  1. feat: run.sh

    richard.lin
    29b89a8

  2. feat: unstrict evaluation

    richard.lin
    ec09b0e

  3. feat: sync for offline environment.

    richard.lin
    4bb8584

  1. update: add UniDriveVLA deps into prepare_env.sh

    richard.lin
    a538003

  2. fix: load whole model on single gpu to avoid quant error.

    richard.lin
    f38cc27

  3. fix: remove ONNX_NVFP4 model for UniDriveVLA_Nusc_Base_Stage1 model.

    richard.lin
    9403cf0

  1. fix: GPT-score can't work in China.

    richard.lin
    f663d2c

  1. add: unidrive_vla_nusc_base_evaluation

    richard.lin
    ecfbb2c

  2. fix: finetune output folder.

    richard.lin
    620f331

  1. feat: add qat functions.

    richard.lin
    98ff1f3

  1. update: log output.

    richard.lin
    65012cb

  1. feat: add qat for hgnetv2

    richard.lin
    f0b2d36

  2. fix: exclude QKV, attention, MatMuls, LayerNorm, MatMuls layers for int8 quant.

    richard.lin
    570b630

  3. add: mode selector for eval_quantized script.

    richard.lin
    67c0f59

  4. add: vit_large_patch16_224 quant & eval feature

    richard.lin
    28b5253

  1. feat: add prepare cloud env script. optimize script for limit network env.

    richard.lin
    cb154ba

  1. feat: finished with regnet_x_800mf PTQ scripts.

    richard.lin
    91075d7

  2. feat: finished with hgnetv2 PTQ scripts.

    richard.lin
    0678520

  1. feat: save engine and calibration function, model quantize function.

    richard.lin
    5f63376

  1. changed the evaluation datasets.

    richard.lin
    e46bcd9

  2. fix: gradio lite index.html

    richard.lin
    e54f725

  3. add: model infer, model eval, within HuggingFace APIs, hello

    richard.lin
    db9ebae

  1. initial commit

    richardlin823
    Verified signature
    02b6610

  2. Duplicate from gradio/gradio-lite-template⁠

    richardlin823 +1: Xenova
    414b617