https://github.com/sanyam8055/sentiment_analyser

https://github.com/sanyam8055/sentiment_analyser

Last synced: 30 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/sanyam8055/sentiment_analyser
• Owner: Sanyam8055
• Created: 2024-05-20T12:42:36.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-05-20T15:14:20.000Z (8 months ago)
• Last Synced: 2024-05-20T17:50:22.789Z (8 months ago)
• Language: Jupyter Notebook
• Size: 74.2 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
### Preprocessing and Training

The notebooks have all the preprocessing, dataset analysis, and training of models along with saving weights. 

There are 4 notebooks

1. notebook_without_training_cleanup - This notebook takes in the dataset without the neutral class and trains models on top with some analysis.

2. notebook_with_common_words_based_cleanup - This notebook analyses common words and their presence in the dataset then trains the model on top.

3. notebook_with_training_data_filtering - This notebook dives deeper into the presence of neutral sentences and then trains the model.

4. notebook_with_bert_glove_embeddings - This notebook trains the Bert model along with glove embeddings. 

Best Performance with the models - notebook_with_training_data_filtering

### The trained model API is live 

curl -X POST "http://20.42.62.249:8082/internal/sentiment_analysis/analyse/text" \        

  -H "Accept: */*" \

  -H "Content-Type: application/json" \

  -d '{"text":"This is such a bad day"}'

Expected Output 

{

  "sentiment": "NEGATIVE",

  "status": 1

}

### The swagger code is entirely Dockerized