dabboud/Commgpt-3B
Full fine-tune vs unsloth/Qwen2.5-3B-Instruct. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 59/73, MLP 108/108, Attention 250/252
2 derivatives checked
- mradermacher/Commgpt-3B-i1-GGUF: Quantization
- mradermacher/Commgpt-3B-GGUF: Same weights, converted