mergekit-community/dolphin-mistral-instruct-7b
Full fine-tune vs arcee-ai/sec-mistral-7b-instruct-1.6-epoch. Fine-tune: 97% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 20/65, MLP 93/96, Attention 124/128, lm_head 1/1
1 derivatives checked
- MaziyarPanahi/dolphin-mistral-instruct-7b-GGUF: Quantization
10 other repos hold byte-for-byte identical weight files.