mistralai/Mixtral-8x22B-Instruct-v0.1
Full fine-tune vs mistralai/Mixtral-8x22B-v0.1. Fine-tune: 100% of the weights changed, across 56 of 56 layers.
Tensors changed: MLP 1344/1344, other 56/56, Attention 224/224, Norms 87/113
2 derivatives checked
- mradermacher/Mixtral-8x22B-Instruct-v0.1-i1-GGUF: Quantization
- MaziyarPanahi/Mixtral-8x22B-Instruct-v0.1-AWQ: Quantization
7 other repos hold byte-for-byte identical weight files.