FastVideo/FastVideo-FastH3-8-Step-V2
Full fine-tune vs MiniMaxAI/MiniMax-H3. Fine-tune: 99% of the weights changed, across 0 of 0 layers.
Tensors changed: Attention 212/212, other 207/208, Embeddings 6/6, Norms 18/212
4 derivatives checked
- realrebelai/FastH3-V2_GGUFs: Not comparable yet
- FastVideo/FastVideo-FastH3-8-Step-V2-NVFP4-Consumer: Same weights, repacked (fewer layers)
- FastVideo/FastVideo-FastH3-Trim-8-Step-NVFP4: Full fine-tune
- FastVideo/FastVideo-FastH3-Trim-8-Step: Full fine-tune