Instructions to use lvladikov/SeedVR2-1.4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- SeedVR
How to use lvladikov/SeedVR2-1.4B with SeedVR:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- MLX
How to use lvladikov/SeedVR2-1.4B with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir SeedVR2-1.4B lvladikov/SeedVR2-1.4B
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
Has anyone compared this with seedvr2 3b?
Has anyone compared it with seedvr2_3b? It would be great if the effect is better than seedvr2_3b.
I am running detailed analysis and would provide details later.
I have done full analysis and comparison with 3B, and it is honest very detailed one - 3B (being bigger 2x many params and 6x more layers does win, but my model has it's place too, as it was created to be the smallest descent quality version of the amazing SeedVR2, and I am personally very happy with it, and I use it on my phone and tablet and does a speedy and good job there).
Full detailed comparison with 3B including same testing images on both my model and the 3B one: https://huggingface.co/lvladikov/SeedVR2-1.4B#compared-to-the-official-seedvr2-3b
Feel free to compare same prompt/seeds with both. Don't just look at them at 100% crop zoom :) look at them as a whole. my 1.4B does a good job, but if you are after fidelity only and the size and speed and memory requirements don't bother you, then go with the official 3b/7b models.
https://huggingface.co/lvladikov/SeedVR2-1.4B/tree/main/assets/official-3B
https://huggingface.co/lvladikov/SeedVR2-1.4B/tree/main/assets/student-1.4B