Update README.md
Browse files
README.md
CHANGED
|
@@ -29,3 +29,10 @@ configs:
|
|
| 29 |
- split: train
|
| 30 |
path: data/train-*
|
| 31 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
- split: train
|
| 30 |
path: data/train-*
|
| 31 |
---
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
(Unofficial) Mini version of Dynamath (https://huggingface.co/datasets/DynaMath/DynaMath_Sample) dataset.
|
| 35 |
+
|
| 36 |
+
Here, we have sampled 100 random ids from all 10 variants of the original Dynamath dataset.
|
| 37 |
+
|
| 38 |
+
This allows faster inference on 1000 eval examples instead of 5010 examples; which can be very slow and limiting for large model runs.
|