jsmith0475/Qwen2.5-7B-bnb-4bit-smith-neuroscience
Full fine-tune vs unsloth/Qwen2.5-7B. jsmith0475/Qwen2.5-7B-bnb-4bit-smith-neuroscience is a GGUF (Q4_K_M) of different weights. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 0/57, MLP 84/84, Attention 112/196, lm_head 1/1