sdworld/Qwen3.8-Next-SDnvfp2

Partial fine-tune vs RadixArk/Qwen3.8-Flash-Next-NVFP4. Partial change: 0% of the weights changed (mlp), in 48 of 48 layers.

Tensors changed: MLP 728/1589, Embeddings 0/138, Norms 0/302, other 0/8, Attention 0/608, lm_head 0/1

On the Hub · Downloads on Model Pulse