NexaAI/Octopus-v2
Full fine-tune vs google/gemma-2b. Fine-tune: 100% of the weights changed, across 18 of 18 layers.
Tensors changed: Norms 37/37, MLP 54/54, Attention 72/72
5 derivatives checked
- second-state/Octopus-v2-GGUF: Same weights, converted
- mradermacher/Octopus-v2-i1-GGUF: Quantization
- DevQuasar-8/NexaAIDev.Octopus-v2-GGUF: Same weights, converted
- mradermacher/Octopus-v2-GGUF: Same weights, converted
- tensorblock/Octopus-v2-GGUF: Quantization