rodrigomt/gama-12b

Full fine-tune vs allura-org/Gemma-3-Glitter-12B. Fine-tune: 97% of the weights changed, across 48 of 48 layers.

Tensors changed: Embeddings 1/4, Norms 285/400, MLP 150/252, Attention 197/408, other 1/1

2 derivatives checked

• mradermacher/gama-12b-i1-GGUF: Quantization (fewer layers)
• mradermacher/gama-12b-GGUF: Quantization (fewer layers)

On the Hub · Downloads on Model Pulse