Sentiment Analysis Run and Submit Tutorial
Sentiment Analysis by INT3405E_41_DMT (Duc Tran Van, Minh Tan Nguyen, Trung Hoang Nguyen). Rank first in class INT3405E 41
- Firstly, download data.zip file above then extract and then open int3405-phobert-vgg11.ipynb config the cell like this, {BASE_DIR} is your current dataset directory you have extracted.
DATASET_TRAIN_TEXT_FILE = '{BASE_DIR}/machine-learning-preproc/train_merged_segmented_v2.csv'
DATASET_TRAIN_IMAGE_FOLDER = '{BASE_DIR}/int3405-ml-preprocess-image-data-v2/image_train_resized'
DATASET_TEST_TEXT_FILE = '{BASE_DIR}/machine-learning-preproc/test_labelled_segmented.csv'
DATASET_TEST_IMAGE_FOLDER = '{BASE_DIR}/int3405-ml-preprocess-image-data-v2/image_test_resized'- Secondly, install python library file in requirements.txt if run in local, otherwise Colab or Kaggle, just skip this step.
- Thirdly, make sure internet is connected for model downloading and GPU is activated for model training.
- Fourthly, run all the cell in int3405-phobert-vgg11.ipynb and wait for training.
- Finally, we use submission.csv file to submit to INT3405 - Sentiment Analysis Problem
review_d.to_csv("submission.csv", index=False)