| license: mit | |
| datasets: | |
| - mlabonne/guanaco-llama2 | |
| # Finetuning Llama2 7b on 1000 samples from the Guanaco dataset, using QLoRA - quantization + low rank adaptation. | |
| The Llama2 model from Meta has several variants. The smallest one containing 7 billion parameters has been finetuned here. |