Info

This is a heavily pruned version of Llama 3.1 8B Instruct which went from 32 layers, all the way down to 9. RLHF is definitely required, and is in the works. Hopefully all goes well and we have a Mini Llama 3.1!

Uploaded model

Developed by: QuietImpostor
License: apache-2.0
Finetuned from model : QuietImpostor/Llama-3.1-Mini-Instruct

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

Downloads last month: 67

Safetensors

Model size

3.01B params

Tensor type

FP16

Inference Providers NEW

Text Generation

This model is not currently available via any of the supported Inference Providers.

Model tree for QuietImpostor/Llama-31-Mini-Instruct-ft-fp16

Base model

QuietImpostor/Llama-3.1-Mini-Instruct

Finetuned

(1)

this model