sail/Qwen2.5-Math-7B-Oat-Zero

Full fine-tune vs Qwen/Qwen2.5-Math-7B. Fine-tune: 92% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 0/1, Norms 54/57, MLP 83/84, Attention 196/196, lm_head 1/1

1 derivatives checked

• mradermacher/Qwen2.5-Math-7B-Oat-Zero-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse