glogwa68/granite-4.0-h-350m-DISTILL-gemini-think

Full fine-tune vs ibm-granite/granite-4.0-h-350m. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/1, Norms 30/93, other 177/196, MLP 64/64, Attention 16/16

1 derivatives checked

• glogwa68/granite-4.0-h-350m-DISTILL-gemini-think-GGUF: Not comparable yet

On the Hub · Downloads on Model Pulse