search

Sherrylone / PQDiff

[ICLR 2024] Continuous-Multiple Image Outpainting in One-Step via Positional Query and A Diffusion-based Approach Link: https://arxiv.org/abs/2401.15652
65 stars 1 forks source link

Question #12

Open ucasyjz opened 2 months ago

ucasyjz commented 2 months ago

when we load the weight of scenery, we meet the error RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory can you give me some guidance, thank you

ayushvrma commented 1 month ago

i think there is some issue in them uploading the weights that makes them corrupt. Please fix @Sherrylone

Sherrylone commented 2 weeks ago

Thank you for your attention to our work. However, as I cannot access the original server, I currently don't have sufficient computational resources to perform retraining. If you have adequate computational resources (8 V100 GPUs or 4 A100 GPUs), you can run the training script to reproduce the results.

ustczhouyu commented 2 weeks ago

Thank you for your attention to our work. However, as I cannot access the original server, I currently don't have sufficient computational resources to perform retraining. If you have adequate computational resources (8 V100 GPUs or 4 A100 GPUs), you can run the training script to reproduce the results.

Why are the Sampling Results by the Provided Pre-trained Model and My model trained by myself Colorful and Chaotic? My Sampling Command Is CUDA_VISIBLE_DEVICES=0 python evaluate.py --target_expansion 0.25 0.25 0.25 0.25 --eval_dir. /eval_dir/scenery/2x/ --size 128 --config flickr192_large. Did anyone get the correct sampling results?