anakin87/gemma-2b-orpo
Full fine-tune vs google/gemma-2b. Fine-tune: 100% of the weights changed, across 18 of 18 layers.
Tensors changed: Embeddings 1/1, Norms 32/37, MLP 54/54, Attention 72/72
4 derivatives checked
- mradermacher/gemma-2b-orpo-i1-GGUF: Quantization
- Felladrin/gguf-sharded-gemma-2b-orpo: Quantization
- mradermacher/gemma-2b-orpo-GGUF: Same weights, converted
- Felladrin/gguf-gemma-2b-orpo: Same weights, converted