richard.lin
feat: run.sh
29b89a8 - PLANS add: unidrive_vla_nusc_base_evaluation
- __pycache__ add: model infer, model eval, within HuggingFace APIs, hello
- 1.52 kB Duplicate from gradio/gradio-lite-template
- 1.97 kB add: model infer, model eval, within HuggingFace APIs, hello
- 13.9 kB fix: GPT-score can't work in China.
- 228 Bytes add: model infer, model eval, within HuggingFace APIs, hello
- 12.4 kB update: log output.
- 18 kB update: log output.
- 33.9 kB fix: finetune output folder.
- 25.7 kB fix: finetune output folder.
- 6.88 kB feat: finished with hgnetv2 PTQ scripts.
- 8.96 kB update: add UniDriveVLA deps into prepare_env.sh
- 12.4 kB update: log output.
- 17.9 kB update: log output.
- 6.91 kB feat: finished with regnet_x_800mf PTQ scripts.
- 21.2 kB feat: finished with regnet_x_800mf PTQ scripts.
- 22.7 kB feat: sync for offline environment.
- 7.58 kB feat: finished with regnet_x_800mf PTQ scripts.
- 740 Bytes feat: run.sh
- 45.6 kB feat: unstrict evaluation
- 39.7 kB fix: load whole model on single gpu to avoid quant error.
- 16 kB update: log output.
- 18.1 kB update: log output.
- 18 kB fix: exclude QKV, attention, MatMuls, LayerNorm, MatMuls layers for int8 quant.