| File List for IFSC7370 Data Science Course, Spring, 2024 | |
| (1) BiaoFu-IFSC7370-DS-SP24.pdf | |
| This is the report file for the assignment. | |
| I do the analysis by developing some app using Hugging Face. | |
| The URL of the Hugging Face app is listed in the report. | |
| (2) 3MT-BiaoFu.ppt | |
| This is the 3 minute presentation, the audio is inserted to the ppt file. | |
| (3) The python codes are printed to the report, the original codes are in my spaces. | |
| You can review the full code under each space, and the report includes the URL for each space. | |
| (4) Folder: pythoncode in the raw date and results.zip file. | |
| This is the folder to store some of the codes downloaded from my Hugging Face app. | |
| app-facefeature-w1-full working.py #this file can output excel with image, more details in report. | |
| app-facefeaturefast-w1-full working.py #this file can run faster than the file above, no excel output, more details in report. | |
| app-facebasic-w1-full working.py #this file can display image and emotion score side by side, more details in report. | |
| (5) raw date and results.zip | |
| This is the raw data and app output for doing the report and data analysis. | |
| For example: | |
| (a) emotion_scores (12)-for project.xlsx | |
| This is the excel file output from the app for doing analysis for the report. | |
| The app embed the image to cell for easy datamining, etc. | |
| The excel table is scalable according to the file quantity, though the file has 10 samples as far. | |
| For example, the table can have 100 images if we have 100 images. | |
| (b) For the emotion class, I make two folders initially to store the raw data and result. | |
| The FER2013 dataset has more than 30,000 images, I sampled 100 images for each class for the initial analysis. | |
| The FER2013 dataset was downloaded from the URL: https://www.kaggle.com/datasets/msambare/fer2013?resource=download. | |
| The Hugging Face I developed can do full dataset, however, I need to use the pay version for using the GPU. | |
| For example, once I can get the sponsors, I can re-run the data in a larger or full scale. | |
| For example, the folder name starts with an is for "angry" class. | |
| an: angry | |
| ds: disgust | |
| fe: fear | |
| hp: happy | |
| ne: neutral | |
| sd: sad | |
| sp: surprise | |
| For the initial two folders belonged to angry, | |
| antd: this folder is for store the train data and test data, | |
| where the subfolder ant100 for test data and antr100 for train data, | |
| ant-re: this folder is for store the *.CSV file, chart file (pie, bar, heatmap) output from the test data, | |
| while the subfolder tra is for *.CSV file, chart file (pie, bar, heatmap) output from the train data. | |
| the other class raw data and result are stored in the related folder following the similar naming style above. | |
| (c) Folder name: sample image | |
| This is the folder for storing the sample image for the report. | |