Keras for Text Classification using Vertex AI Reviews

Could not complete the lab. Stuck at the following step, by timeout error: "On the Notebook instances page, click New Notebook > TensorFlow Enterprise > TensorFlow Enterprise 2.6 (with LTS) > Without GPUs."

Oleksandr P. · Reviewed about 2 years ago

Sergio R. · Reviewed about 2 years ago

Dipesh k. · Reviewed about 2 years ago

Richard A. · Reviewed about 2 years ago

Arun M. · Reviewed about 2 years ago

Great lab, outdated information on the BQ dataset though. Table stories no longer exist. Had to use full table and add extra where condition to limit to stories Additionally the where condition to exclude malformed urls wasn't working. Had to make my own to avoid index out of bound

Brett L. · Reviewed about 2 years ago

Volodymyr S. · Reviewed about 2 years ago

Arun P. · Reviewed about 2 years ago

Vimal K. · Reviewed about 2 years ago

Rushi S. · Reviewed about 2 years ago

Looks like the dataset is `bigquery-public-data.hacker_news.full` and not `bigquery-public-data.hacker_news.stories`

Nick G. · Reviewed about 2 years ago

Notebook does not work

Jia Jun G. · Reviewed about 2 years ago

Luigi B. · Reviewed about 2 years ago

Avneesh D. · Reviewed about 2 years ago

Pramod-Kumar A. · Reviewed about 2 years ago

Shalika A. · Reviewed about 2 years ago

aishani G. · Reviewed about 2 years ago

some areas require checking code safe_offset is required

Cesar D. · Reviewed about 2 years ago

Halasi G. · Reviewed about 2 years ago

Nikhil B. · Reviewed about 2 years ago

SANDEEP G. · Reviewed about 2 years ago

There are issues mentioned in other comments. But after fixing them it seems to work. However it doesn't do anything after training the 3 models. Also plots don't work for some reason

Jaroslaw F. · Reviewed about 2 years ago

Roman S. · Reviewed about 2 years ago

The dataset didnt exist. I had go to big quert to find the path. The path was "bigquery-public-data.hacker_news.full" instead of "bigquery-public-data.hacker_news.stories". Also, the regex was not working properly to obtain source. I had to change the regex mentioned in code. The dataset was highly biased and contained only "github" urls and very less "nytimes" urls.

Abhishek V. · Reviewed about 2 years ago

Mustafa Cihad G. · Reviewed about 2 years ago