jongstra/Swift-1.5-Qwen3.8-27b-oQ4e-fp16-mtp

Same weights, converted vs yottle/Swift-1.5-Qwen3.8-27b-oQ4e-mtp. Same weights stored in another dtype (U32/BF16 → U32/F16).

Tensors changed: Embeddings 0/7, Norms 0/325, Attention 0/1176, MLP 0/697, lm_head 0/3, other 0/1

On the Hub · Downloads on Model Pulse