spow12/ChatWaifu_72B_v2.2
Full fine-tune vs AXCXEPT/EZO-Qwen2.5-72B-Instruct. Fine-tune: 100% of the weights changed, across 80 of 80 layers.
Tensors changed: Norms 120/161, MLP 240/240, Attention 560/560
1 derivatives checked
- mradermacher/ChatWaifu_72B_v2.2-i1-GGUF: Quantization