search

ali-vilab / VGen

Official repo for VGen: a holistic video generation ecosystem for video generation building on diffusion models
https://i2vgen-xl.github.io
2.87k stars 254 forks source link

Where is 'workspace/model_bk/model_scope_0267000.pth' in the configs/t2v_train.yaml? #41

Open tacit0428 opened 8 months ago

tacit0428 commented 8 months ago

Pretrain: { 'type': pretrain_specific_strategies, 'fix_weight': False, 'grad_scale': 0.5, 'resume_checkpoint': 'workspace/model_bk/model_scope_0267000.pth', 'sd_keys_path': 'models/stable_diffusion_image_key_temporal_attention_x1.json', }

The Pretrain term has a resume checkpoint. But I can't find the model weight anywhere. Is it this one? https://modelscope.cn/models/damo/text-to-video-synthesis/files/text2video_pytorch_model.pth

Steven-SWZhang commented 8 months ago

Hi, please refer to here.

justinday123 commented 7 months ago

do i have to rename 'text2video_pytorch_model.pth' to 'model_scope_0267000.pth' and change the path?

zhenglinpan commented 4 months ago

do i have to rename 'text2video_pytorch_model.pth' to 'model_scope_0267000.pth' and change the path?

it works for me.