dphn/dolphin-2.9.2-mixtral-8x22b
Partial fine-tune vs mistral-community/Mixtral-8x22B-v0.1. Partial change: 25% of the weights changed (attention, mlp, other), in 54 of 56 layers.
Tensors changed: MLP 336/1344, other 14/56, Attention 56/224, Norms 0/113
1 derivatives checked
- mradermacher/dolphin-2.9.2-mixtral-8x22b-i1-GGUF: Quantization