unsloth/Qwen3.5-4B-MTP-GGUF

Same weights, converted vs Qwen/Qwen3.5-4B. Same weights stored in another dtype (BF16/F32 → BF16/F32), plus 39 new tensors.

Tensors changed: Embeddings 0/1, MLP 0/96, Norms 0/81, Attention 0/32

1 derivatives checked

• shujunyi/Qwen3.5-4B-MTP-Q4_0: Quantization

On the Hub · Downloads on Model Pulse