ernie-research/APPS-Gemma-2B-MA-PPO-Fixed10

Full fine-tune vs google/codegemma-2b. Fine-tune: 100% of the weights changed, across 18 of 18 layers.

Tensors changed: Embeddings 1/1, Norms 37/37, MLP 54/54, Attention 72/72

2 derivatives checked

• mradermacher/APPS-Gemma-2B-MA-PPO-Fixed10-i1-GGUF: Quantization
• mradermacher/APPS-Gemma-2B-MA-PPO-Fixed10-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse