self-media-generation

Dependencies

• python 3.8
• pytorch 2.01
• opencv-python 4.8.0.76
• spack 3.6.1

Workflow

How to run:

1. Select 4 or 9 portrait photos you like.
2. Run the Main.py
3. You will get the processed photo like the third photo in the workflow and the vivid text with emoji😀.
4. Feel free to upload the content to Twitter or any social media platform.

Training

The new bert will be updating soon

Attention

We use a program I that designed(Accessing Mirror Sites with Crawlers) to enter a fake chatgpt to achive call LLM

Validation

In image caption, we use clip to compute a similarity between image and word to control the keyword generating.

Examples in Twitter

Sentiment Analysis and Topic Categorization

• Sentiment Analysis Net is on ./sentimentAnalysisNet, follow the instruction on the readme in the folder to train the net.
• Topic Categorization Net is on ./Bert_classification, follow the instruction on the readme in the folder to train the net.