Update README.md
Browse files
README.md
CHANGED
|
@@ -16,28 +16,3 @@ Unlike version 1 this model has no issues at fp16 or any quantizations.
|
|
| 16 |
The model that was used to create this one is linked below:
|
| 17 |
|
| 18 |
https://huggingface.co/meta-llama/Meta-Llama-3-8B
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
- Llama-3-11.5B-V2
|
| 22 |
-
|
| 23 |
-
| Metric | Value |
|
| 24 |
-
|---------------------------------|------:|
|
| 25 |
-
| Avg. | 66.89 |
|
| 26 |
-
| AI2 Reasoning Challenge(25-Shot)| 57.68 |
|
| 27 |
-
| HellaSwag (10-Shot) | 78.59 |
|
| 28 |
-
| MMLU (5-Shot) | 65.39 |
|
| 29 |
-
| TruthfulQA (0-shot) | 35.86 |
|
| 30 |
-
| Winogrande (5-shot) | 74.74 |
|
| 31 |
-
| GSM8k (5-shot) | 69.37 |
|
| 32 |
-
|
| 33 |
-
- Original Meta-Llama-3-8B
|
| 34 |
-
|
| 35 |
-
| Metric |Value|
|
| 36 |
-
|---------------------------------|----:|
|
| 37 |
-
|Avg. |62.87|
|
| 38 |
-
|AI2 Reasoning Challenge (25-Shot)|59.47|
|
| 39 |
-
|HellaSwag (10-Shot) |82.09|
|
| 40 |
-
|MMLU (5-Shot) |66.69|
|
| 41 |
-
|TruthfulQA (0-shot) |43.90|
|
| 42 |
-
|Winogrande (5-shot) |77.35|
|
| 43 |
-
|GSM8k (5-shot) |45.34|
|
|
|
|
| 16 |
The model that was used to create this one is linked below:
|
| 17 |
|
| 18 |
https://huggingface.co/meta-llama/Meta-Llama-3-8B
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|