EPFLiGHT/Gemma-3-27B-MeditronFO

Full fine-tune vs google/gemma-3-27b-it. Fine-tune: 99% of the weights changed, across 62 of 62 layers.

Tensors changed: Embeddings 1/1, Norms 53/373, MLP 183/186, Attention 247/248

2 derivatives checked

• mradermacher/Gemma-3-27B-MeditronFO-i1-GGUF: Quantization (fewer layers)
• mradermacher/Gemma-3-27B-MeditronFO-GGUF: Quantization (fewer layers)

On the Hub · Downloads on Model Pulse