Add results preview
Browse files
README.md
CHANGED
|
@@ -53,6 +53,29 @@ print(output_text)
|
|
| 53 |
|
| 54 |
Minitron is released under the [NVIDIA Open Model License Agreement](https://developer.download.nvidia.com/licenses/nvidia-open-model-license-agreement-june-2024.pdf).
|
| 55 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
## Citation
|
| 57 |
|
| 58 |
If you find our work helpful, please consider citing our paper:
|
|
|
|
| 53 |
|
| 54 |
Minitron is released under the [NVIDIA Open Model License Agreement](https://developer.download.nvidia.com/licenses/nvidia-open-model-license-agreement-june-2024.pdf).
|
| 55 |
|
| 56 |
+
## Evaluation Results
|
| 57 |
+
|
| 58 |
+
*5-shot performance.* Language Understanding evaluated using [Massive Multitask Language Understanding](https://arxiv.org/abs/2009.03300):
|
| 59 |
+
|
| 60 |
+
| Average |
|
| 61 |
+
| :---- |
|
| 62 |
+
| 58.6 |
|
| 63 |
+
|
| 64 |
+
*Zero-shot performance.* Evaluated using select datasets from the [LM Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) with additions:
|
| 65 |
+
|
| 66 |
+
| HellaSwag | Winogrande | GSM8K| ARC-C | XLSum |
|
| 67 |
+
| :------------- | :------------- | :------------- | :------------- | :------------- |
|
| 68 |
+
| 75.0 | 74.0 | 24.1 | 50.9 | 29.5
|
| 69 |
+
|
| 70 |
+
|
| 71 |
+
*Code generation performance*. Evaluated using [HumanEval](https://github.com/openai/human-eval):
|
| 72 |
+
|
| 73 |
+
| p@1, 0-Shot |
|
| 74 |
+
| :------------- |
|
| 75 |
+
| 23.3 |
|
| 76 |
+
|
| 77 |
+
Please refer to our [paper](https://arxiv.org/abs/2407.14679) for the full set of results.
|
| 78 |
+
|
| 79 |
## Citation
|
| 80 |
|
| 81 |
If you find our work helpful, please consider citing our paper:
|