pyrodog/DeepSeek-V4.1-Flash-UNCENSORED-DwarfStar-Q2

Same weights, repacked (fewer layers) vs dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8. pyrodog/DeepSeek-V4.1-Flash-UNCENSORED-DwarfStar-Q2 matches part of dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8 (266 of its 96085 tensors; the rest of dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8 has no counterpart). On the shared part: Same weights in different files (re-saved, re-sharded or new metadata).

Tensors changed: other 0/7, Attention 0/128, MLP 0/64, Norms 0/65, Embeddings 0/2

On the Hub · Downloads on Model Pulse