Update README.md
Browse files
README.md
CHANGED
|
@@ -12,6 +12,13 @@ tags:
|
|
| 12 |
|
| 13 |
This is a passthrough model based on arco and danube as a first effort to train a small enough reasoning language model that generalizes across all kind of commonsense. The model is untrained.
|
| 14 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
#### Configuration
|
| 16 |
|
| 17 |
The following YAML configuration was used to produce this model:
|
|
|
|
| 12 |
|
| 13 |
This is a passthrough model based on arco and danube as a first effort to train a small enough reasoning language model that generalizes across all kind of commonsense. The model is untrained.
|
| 14 |
|
| 15 |
+
#### Benchmarks
|
| 16 |
+
|
| 17 |
+
| Parameters | Model | MMLU | ARC | HellaSwag | PIQA | Winogrande | Average |
|
| 18 |
+
| -----------|--------------------------------|-------|-------|-----------|--------|------------|---------|
|
| 19 |
+
| 488m | arco-lite | **23.22** | 33.45 | 56.55| 69.70 | **59.19**| 48.46 |
|
| 20 |
+
| 770m | arco-plus | 23.06 | **36.43** | **60.09**|**72.36**| **60.46**| **50.48** |
|
| 21 |
+
|
| 22 |
#### Configuration
|
| 23 |
|
| 24 |
The following YAML configuration was used to produce this model:
|