second-state/Qwen2-72B-Instruct-GGUF

Same weights, converted vs Qwen/Qwen2-72B-Instruct. Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, MLP 0/240, Attention 0/560, Norms 0/161, lm_head 0/1

On the Hub · Downloads on Model Pulse