elinas/Llama-3-13B-Instruct-ft

Full fine-tune vs elinas/Llama-3-13B-Instruct. Fine-tune: 96% of the weights changed, across 55 of 55 layers.

Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 0/111, MLP 165/165, Attention 220/220

1 derivatives checked

• mradermacher/Llama-3-13B-Instruct-ft-GGUF: Quantization

On the Hub · Downloads on Model Pulse