BerenMillidge
commited on
Commit
•
a3433d1
1
Parent(s):
9270cee
Update README.md
Browse files
README.md
CHANGED
@@ -64,10 +64,10 @@ Zyda also outperforms Dolma, RefinedWeb, and Fineweb on 1.4B models trained on 5
|
|
64 |
|
65 |
According to our evaluations, Zyda is the most performant per-token open dataset available in its non-starcoder variant on language tasks and tying with fineweb otherwise.
|
66 |
|
67 |
-
![image/png](https://cdn-uploads.huggingface.co/production/uploads/65c05e75c084467acab2f84a/fXaQAOBDJpoaAr1clfTel.png)
|
68 |
|
69 |
<center>
|
70 |
-
<img src="https://cdn-uploads.huggingface.co/production/uploads/65c05e75c084467acab2f84a/fXaQAOBDJpoaAr1clfTel.png" width="
|
71 |
</center>
|
72 |
|
73 |
These results are an aggregate scores of classic language modelling evaluations (piqa, winogrande, openbookqa, arc-easy, arc-challenge) across time for a 1.4B model trained on 50B tokens of each dataset.
|
|
|
64 |
|
65 |
According to our evaluations, Zyda is the most performant per-token open dataset available in its non-starcoder variant on language tasks and tying with fineweb otherwise.
|
66 |
|
67 |
+
<!-- ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65c05e75c084467acab2f84a/fXaQAOBDJpoaAr1clfTel.png) -->
|
68 |
|
69 |
<center>
|
70 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/65c05e75c084467acab2f84a/fXaQAOBDJpoaAr1clfTel.png" width="1200" alt="Zyda performance across steps.">
|
71 |
</center>
|
72 |
|
73 |
These results are an aggregate scores of classic language modelling evaluations (piqa, winogrande, openbookqa, arc-easy, arc-challenge) across time for a 1.4B model trained on 50B tokens of each dataset.
|