Inquiry regarding pretrained Transformer weights for SAMPO inference
First of all, thank you for sharing the impressive work on SAMPO. I am a researcher interested in video generation and currently trying to reproduce the inference results using your provided predict.py script.
While setting up the environment, I successfully located the VQVAE weights (approx. 456MB) in the SAMPO_VQVAE repository. However, the inference script requires both a tokenizer and a transformer subfolder to function correctly.
Currently, I am unable to find the pretrained Transformer weights (the sequence-level predictor) that match the AutoModelForCausalLM configuration in the script.
Could you please let me know if the Transformer weights are available for public download? If they are hosted in a different repository or if there is a specific way to access them, I would greatly appreciate your guidance.
Thank you for your time and for contributing to the community!