run the Notebook Fizzbuzz task 2.ipynb to deploy the app huggingface-inference on port 8000. the app classifies text as toxic/non-toxic. the last cell in notebook is used to send request to the docker container and receive a response.
the model used is "martin-ha/toxic-comment-model". I chose NLP task because of familiarity, and chose this model because of large number of downloads while also being a reasonable size to deploy. other models i tried were too big and were running into time out errors, so for purpose of demonstration only, i went for a smaller model instead.
3) EXPLORATORY DATA ANALYSIS DEMONSTRATION
for this task I chose the olivierdehaene/xkcd dataset because I wanted to try out multi-modal dataset. again one reason for picking this dataset is because of its smaller size, that was quick to load compared to some of the other datsets.
use the notebook Dataset exploration.ipynb for this task.
the first section shows a wordcloud of the most common words in the title. you can then use these to explore comics by keyword or create a story of multi-part comics.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent