Update README.md
Browse files
README.md
CHANGED
|
@@ -67,8 +67,7 @@ It achieves the following results on the evaluation set:
|
|
| 67 |
While the macro F1 scores across the 1-5 rating scale are relatively low due to the model's difficulty in distinguishing between higher-rated samples, the classifier performs well for our primary filtering task. When converting to binary classification, using a threshold of 2 achieves the F1 scores ranges between 0.8 and 0.9 for most Stack-Edu classifiers, whereas a threshold of 3 yields F1 scores between 0.5 and 0.8. With the Highest being Python, SQL, C, Rust and the lowest being TypeScript and PHP.
|
| 68 |
|
| 69 |
<div style="display: flex; justify-content: center; gap: 20px;">
|
| 70 |
-
<img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/VYh1deFc8Jif7B4kDpndc.png" width="
|
| 71 |
-
<img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/I_O96PgYAZ9pj27lM4B9a.png" width="500">
|
| 72 |
</div>
|
| 73 |
|
| 74 |
We validated these classifiers by filtering Stack v2 data and testing on an intermediate SmolLM2 checkpoint. Filtering with a threshold of 3 improved performance across most languages while maintaining adequate data volume, though Java showed better results with a threshold of 2.
|
|
|
|
| 67 |
While the macro F1 scores across the 1-5 rating scale are relatively low due to the model's difficulty in distinguishing between higher-rated samples, the classifier performs well for our primary filtering task. When converting to binary classification, using a threshold of 2 achieves the F1 scores ranges between 0.8 and 0.9 for most Stack-Edu classifiers, whereas a threshold of 3 yields F1 scores between 0.5 and 0.8. With the Highest being Python, SQL, C, Rust and the lowest being TypeScript and PHP.
|
| 68 |
|
| 69 |
<div style="display: flex; justify-content: center; gap: 20px;">
|
| 70 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/VYh1deFc8Jif7B4kDpndc.png" width="600">
|
|
|
|
| 71 |
</div>
|
| 72 |
|
| 73 |
We validated these classifiers by filtering Stack v2 data and testing on an intermediate SmolLM2 checkpoint. Filtering with a threshold of 3 improved performance across most languages while maintaining adequate data volume, though Java showed better results with a threshold of 2.
|