|
--- |
|
tags: |
|
- Tutorial |
|
size_categories: |
|
- n<1K |
|
--- |
|
|
|
## Zero to One: Label Studio Tutorial Dataset |
|
|
|
This dataset is used in the [Label Studio Zero to One Tutorial](https://hubs.ly/Q01CNlyy0). This dataset was originally provided by [Andrew Maas](https://ai.stanford.edu/~amaas/)([ref](https://ai.stanford.edu/~amaas/papers/wvSent_acl2011.bib)). This is an open and well-known dataset. The original dataset did have over 100,000 reviews. |
|
|
|
### Parsing down 100,000 reviews to 100 reviews |
|
|
|
To parse this dataset down to 100 reviews, (Chris Hoge)[https://huggingface.co/hogepodge] and myself((Erin Mikail Staples)[https://huggingface.co/erinmikail]) took the following steps. |
|
|
|
It started by (writing a script)[https://s3.amazonaws.com/labelstud.io/datasets/IMDB_collect.py] that walked the directory structure to capture the data and metadata as rows of data. The data was written in randomized batches with rows corresponding to: |
|
|
|
- 0 - 25,000: Labeled training data, with positive and negative sentiment mixed. |
|
- 25,001 - 75000: Unlabeled training data. |
|
- 75001 - 100,000: Labeled testing data, with positive and negative sentiment mixed. |
|
|
|
These batches were also written out as separate files for convenience. Finally, the first 100 rows of each batch were written out as separate files to support faster loading for a streamlined learning experience. |
|
|
|
Our thanks to Andrew Maas for having provided this free data set from their research. |
|
|
|
## Did you try your hand at this tutorial? |
|
|
|
We'd love to hear you share your results and how it worked out for you! |
|
|
|
Did you build something else with the data? |
|
|
|
Let us know! Join us in the (Label Studio Slack Community)[https://hubs.ly/Q01CNprb0] or drop us an (email)[mailto:[email protected]] |
|
|
|
## Enjoy what we're working on? |
|
|
|
Drop us a star on (GitHub!)[https://hubs.ly/Q01CNp4W0] |
|
|
|
|
