eaddario/DeepSeek-R1-Distill-Llama-8B-GGUF

Same weights, converted vs deepseek-ai/DeepSeek-R1-Distill-Llama-8B. Same weights stored in another dtype (BF16 → F16/F32), plus 1 new tensors.

Tensors changed: Embeddings 0/1, Attention 0/128, MLP 0/96, Norms 0/65, lm_head 0/1

On the Hub · Downloads on Model Pulse