How yukti20/ecomm-sentiment-analysis works
This is a notebook-based demo for fine-tuning a Hugging Face sentiment model on Amazon product reviews and using it for binary sentiment classification. The README documents the dataset fields, pretrained model, and optional saving of fine-tuned outputs. The notebook source was not sampled, so the graph reflects documented capabilities rather than verified internal calls; preprocessing and evaluation details are omitted.
Main entry points: ecomm_sentiment_analysis.ipynb Analysis notebook
Components 7
Review data
- Amazon Polarity
ecomm_sentiment_analysis.ipynb
Sentiment model
- Analysis notebook
ecomm_sentiment_analysis.ipynb - DistilBERT base
ecomm_sentiment_analysis.ipynb - Fine-tune model
ecomm_sentiment_analysis.ipynb
Results
- Classify reviews
ecomm_sentiment_analysis.ipynb - Saved artifacts
ecomm_sentiment_analysis.ipynb
Other components
- Notebook user
Connections 7
- Notebook user → Analysis notebook: runs
- Analysis notebook → Amazon Polarity: loads splits
- Analysis notebook → DistilBERT base: loads model
- Amazon Polarity → Fine-tune model: provides reviews
- DistilBERT base → Fine-tune model: provides weights
- Fine-tune model → Classify reviews: uses tuned model
- Analysis notebook → Saved artifacts: optionally saves