Gorilla4X/Quacken-R1-14B-FP8
Same weights, converted vs unsloth/DeepSeek-R1-Distill-Qwen-14B. Same weights stored in another dtype (BF16 → T43/F16).
Tensors changed: Embeddings 0/1, Norms 0/97, Attention 0/144, lm_head 0/1
Same weights, converted vs unsloth/DeepSeek-R1-Distill-Qwen-14B. Same weights stored in another dtype (BF16 → T43/F16).
Tensors changed: Embeddings 0/1, Norms 0/97, Attention 0/144, lm_head 0/1