|
--- |
|
license: other |
|
license_name: llama-3 |
|
license_link: https://llama.meta.com/llama3/license/ |
|
--- |
|
|
|
Llama-3-11.5B-Instruct-v2 |
|
|
|
Thank you to Meta for the weights for Meta-Llama-3-8B-Instruct |
|
|
|
 |
|
|
|
This is an upscaling of the Meta-Llama-3-8B-Instruct Ai using techniques created for chargoddard/mistral-11b-slimorca. This Ai model has been upscaled from 8b parameters to 11.5b parameters without any continuous pretraining or fine-tuning. |
|
|
|
Unlike version 1 this model has no issues at fp16 or any quantizations. |
|
|
|
The model that was used to create this one is linked below: |
|
|
|
https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct |
|
______________________________________________________________________________ |
|
- Llama-3-11.5B-Instruct-V2 |
|
|
|
| Metric | Value | |
|
|----------------------------------|-------:| |
|
| Avg. | 63.91 | |
|
| AI2 Reasoning Challenge (25-Shot)| 57.68 | |
|
| HellaSwag (10-Shot) | 78.59 | |
|
| MMLU (5-Shot) | 67.35 | |
|
| TruthfulQA (0-shot) | 35.86 | |
|
| Winogrande (5-shot) | 74.74 | |
|
| GSM8k (5-shot) | 69.37 | |
|
|
|
- Original Meta-Llama-3-8B-Instruct |
|
|
|
| Metric | Value | |
|
|----------------------------------|-------:| |
|
| Avg. | 66.87 | |
|
| AI2 Reasoning Challenge (25-Shot)| 60.75 | |
|
| HellaSwag (10-Shot) | 78.55 | |
|
| MMLU (5-Shot) | 67.07 | |
|
| TruthfulQA (0-shot) | 51.65 | |
|
| Winogrande (5-shot) | 74.51 | |
|
| GSM8k (5-shot) | 68.69 | |
