JoPmt/MinLlama-3-8B-instruct-pass
Partial fine-tune (fewer layers) vs NousResearch/Meta-Llama-3-8B-Instruct. JoPmt/MinLlama-3-8B-instruct-pass matches part of NousResearch/Meta-Llama-3-8B-Instruct (201 of its 291 tensors; the rest of NousResearch/Meta-Llama-3-8B-Instruct has no counterpart). On the shared part: Partial change: 52% of the weights changed (attention, mlp, norms), in 14 of 22 layers.
Tensors changed: Embeddings 0/1, Norms 28/45, MLP 42/66, Attention 56/88, lm_head 0/1
1 derivatives checked
- mradermacher/MinLlama-3-8B-instruct-pass-GGUF: Same weights, converted