Datasets:
andreaskoepf
commited on
Commit
•
d9fe14f
1
Parent(s):
261a96b
oasst1 -> oasst2
Browse files
README.md
CHANGED
@@ -371,7 +371,7 @@ While HF datasets is ideal for tabular datasets, it is not a natural fit for nes
|
|
371 |
Nevertheless, we make all messages which can also be found in the file `2023-04-12_oasst_ready.trees.jsonl.gz` available in parquet as train/validation splits.
|
372 |
These are directly loadable by [Huggingface Datasets](https://pypi.org/project/datasets/).
|
373 |
|
374 |
-
To load the
|
375 |
|
376 |
```python
|
377 |
from datasets import load_dataset
|
|
|
371 |
Nevertheless, we make all messages which can also be found in the file `2023-04-12_oasst_ready.trees.jsonl.gz` available in parquet as train/validation splits.
|
372 |
These are directly loadable by [Huggingface Datasets](https://pypi.org/project/datasets/).
|
373 |
|
374 |
+
To load the oasst2 train & validation splits use:
|
375 |
|
376 |
```python
|
377 |
from datasets import load_dataset
|