Preparing to train

[kabachuha]

• Add dataset splitter
• Add a low level clip/image captioner like BLIP2 if the description is not available
• Use an external LLM like OpenAI's ChatGPT or plain GPT-3, or even a local model like LLaMa via APIs to generate descriptions for higher level videos by combining the subclips with a global prompt
• Improve the batching process
• Do a test train

[kabachuha]

Ideas for subclips captioners:

General https://github.com/ExponentialML/Video-BLIP2-Preprocessor

Anime https://github.com/bruefire/WaifuTaggerForVideo