0xSero/Step-3.7-Flash-148B

Same weights, repacked vs stepfun-ai/Step-3.7-Flash-NVFP4. Same weights in different files (re-saved, re-sharded or new metadata).

Tensors changed: Attention 0/518, Embeddings 0/2, Norms 0/392, MLP 0/332, other 0/106, lm_head 0/1

On the Hub · Downloads on Model Pulse