Aashraf995/Qwen-Evo-7B
Full fine-tune vs fblgit/cybertron-v4-qw7B-UNAMGS. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: lm_head 1/1, Embeddings 1/1, Norms 57/57, MLP 84/84, Attention 192/196
2 derivatives checked
- mradermacher/Qwen-Evo-7B-i1-GGUF: Quantization
- mradermacher/Qwen-Evo-7B-GGUF: Same weights, converted