LLM model setup and some more general project improvements.
The current pipeline is working but all of the prompts should still be considered suboptimal. However, the pipeline is quite fast by now (~10 minutes for a 2 min video on colab T4), so experimentation should be feasible.
As the pipeline is set up and the code is working, this PR is ready to merge
Closes #31 #23 #16
LLM model setup and some more general project improvements. The current pipeline is working but all of the prompts should still be considered suboptimal. However, the pipeline is quite fast by now (~10 minutes for a 2 min video on colab T4), so experimentation should be feasible.
As the pipeline is set up and the code is working, this PR is ready to merge