Added notebook for sentiment analysis

[vinaybagade]

No description provided.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[vinaybagade]

cc @VibhuJawa

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:11Z
----------------------------------------------------------------

Line #1.    !pip install hvplot

Might be worth commenting this out to keep the notebook clean in a run.

Also worth adding s3fs. 

Line #3. !!pip install s3fs

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:11Z
----------------------------------------------------------------

This code adds complexity.

I suggest switching to below, Its much faster and only uses cudf . Also might be useful to switch to below, it only takes 30 seconds.

%%time
input_bucket = 's3://amazon-reviews-pds'
input_path = '/parquet/product_category=Office_Products/*.parquet'
df = cudf.read_parquet(input_bucket+input_path,

                       storage_options={'anon': True},

                       columns = ['star_rating','review_body'])

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:12Z
----------------------------------------------------------------

Add a markdown section explaing that you are merging reviews here.

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:13Z
----------------------------------------------------------------

Add a section here explaining that you are gonna be tokenizing here.

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:13Z
----------------------------------------------------------------

Add a section to say you are starting a Data Loader

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:14Z
----------------------------------------------------------------

Line #2.    from transformers import BertModel

Mark Down section explaining that you are creating a model here.

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:14Z
----------------------------------------------------------------

Line #3.    def train_model(model, data_loader, loss_fn, optimizer, scheduler, n_examples):

Add a markdown section explaining training loop. .

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

VibhuJawa commented on 2021-09-01T23:16:15Z
----------------------------------------------------------------

Line #1.    ## Use some custom examples

Add a markdown section with custom examples.

[VibhuJawa]

Hey @vinaybagade . Thanks a lot for working on this. The example is really useful across board.

I have requested some changes which you can view here.