mayank-mishra/granite-3b-code-glaive-20k

Full fine-tune vs ibm-granite/granite-3b-code-base-2k. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/1, Norms 65/65, MLP 192/192, Attention 256/256

1 derivatives checked

• mradermacher/granite-3b-code-glaive-20k-GGUF: Partial fine-tune

On the Hub · Downloads on Model Pulse