bunnycore/QwenMosaic-7B

Full fine-tune vs ATH-MaaS/Marco-o1. Fine-tune: 100% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 1/1, Norms 55/57, MLP 84/84, Attention 196/196, lm_head 1/1

2 derivatives checked

• mradermacher/QwenMosaic-7B-i1-GGUF: Quantization
• mradermacher/QwenMosaic-7B-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse