mnm373/RevLlama-Qwen2.5-Coder-1.5B-40000samples_gguf
Full fine-tune vs unsloth/Qwen2.5-Coder-1.5B-Instruct. mnm373/RevLlama-Qwen2.5-Coder-1.5B-40000samples_gguf is a GGUF (Q8_0) 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