Spaces:
Sleeping
Sleeping
steps=300 | |
guidance=3 | |
num_samples=1 | |
model="vta-ldm-clip4clip-v-large" | |
# model="vta_ldm_clip4clip_augment_pe" | |
# model="vta_ldm_clip4clip_augment_ib" | |
# model="vta_ldm_clip4clip_ib" | |
# model="vta_ldm_clip4clip_lb" | |
# model="vta_ldm_clip4clip_pe" | |
# model="vta_ldm_clip4clip_text" | |
# model="vta_ldm_youtube" | |
# model="vta_ldm_vjepa" | |
CUDA_VISIBLE_DEVICES=2 python3.10 inference_from_video.py --original_args="ckpt/$model/summary.jsonl" \ | |
--model="ckpt/$model/pytorch_model_2.bin" \ | |
--sample_rate 16000 \ | |
--data_path "data" \ | |
--save_dir outputs/$model \ | |
--num_steps $steps \ | |
--guidance $guidance \ | |
--num_samples $num_samples \ | |
--batch_size 8 | |