bartowski/Hermes-2-Pro-Mistral-10.7B-GGUF
Full fine-tune (fewer layers) vs mistralai/Mistral-7B-v0.1. mistralai/Mistral-7B-v0.1 matches part of bartowski/Hermes-2-Pro-Mistral-10.7B-GGUF (289 of its 435 tensors; the rest of bartowski/Hermes-2-Pro-Mistral-10.7B-GGUF has no counterpart). On the shared part: bartowski/Hermes-2-Pro-Mistral-10.7B-GGUF is a GGUF (Q8_0) of different weights. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Norms 28/65, MLP 96/96, Attention 128/128