TarunSinghal/bert-base-uncased-SetimentAnalysis

This fine tunned model will do sentiment analysis,based on 6 sentiments - sadness (0), joy (1), love (2), anger (3), fear (4), surprise (5)

Download and run Colab Notebook "FineTunnedBertBaseModel_Use.ipynb" for step by step guidance, which is in "Files and Versions" section. When you run above google colab file you will require following 3 files which is in "Files and versions" section

1. model.index
2. checkpoint
3. model.data-00000-of-00001 Create a folder on you google drive name, folder name should be "model2.1" and save all these 3 files in it. If you want to change folder name instead of model2.1, you want folder name should be "sentimentXYZ" then you have to change line "classifier_2.load_weights('/content/drive/MyDrive/FineTunning2/model2.1/model')" in "FineTunnedBertBaseModel_Use.ipynb" to "classifier_2.load_weights('/content/drive/MyDrive/FineTunning2/setimentXYZ/model')" this is the path of your save weights on In the following path '/content/drive/MyDrive/FineTunning2/model2.1/model'

Overview BaseModel : "bert-base-uncased" DataSet: dair-ai/emotion

Training: After first epoch, accuracy: 0.6497 After third epoch, accuracy: 0.9360

Test: accuracy: 0.9265

For any contribution or discussion please let me know in Discussion section.