ABrain/Delta-NAS-Mistral-7B
Partial fine-tune vs mistralai/Mistral-7B-Instruct-v0.3. Partial change: 74% of the weights changed (attention, lm_head, mlp), in 24 of 32 layers.
Tensors changed: Embeddings 0/1, Norms 0/65, MLP 72/96, Attention 96/128, lm_head 1/1
2 derivatives checked
- mradermacher/Delta-NAS-Mistral-7B-i1-GGUF: Quantization
- mradermacher/Delta-NAS-Mistral-7B-GGUF: Same weights, converted