https://github.com/bluecube246/youtube-comment-deleter
Created a program to remove YouTube scam comments stating that you have won a prize and to cover shipping expenses on Telegram or Whatsapp.
https://github.com/bluecube246/youtube-comment-deleter
Last synced: 3 months ago
JSON representation
Created a program to remove YouTube scam comments stating that you have won a prize and to cover shipping expenses on Telegram or Whatsapp.
- Host: GitHub
- URL: https://github.com/bluecube246/youtube-comment-deleter
- Owner: bluecube246
- License: apache-2.0
- Created: 2022-08-02T02:45:31.000Z (almost 4 years ago)
- Default Branch: master
- Last Pushed: 2022-09-05T21:45:13.000Z (almost 4 years ago)
- Last Synced: 2025-01-28T12:44:29.548Z (over 1 year ago)
- Language: Python
- Homepage:
- Size: 567 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Youtube Comment Deleter
Created a program to remove YouTube scam comments stating that you have won a prize and to cover shipping expenses on Telegram or Whatsapp.
## Extracting comments from Youtube via channel id.
```
python3 youtube.py --channel_id channelId --n number_of_videos
```
## Label Data
Label all of the spam comments saved in the youtube_data directory to TRUE. All of the comments are by default false.
## Convert labeled data to BIO format
```
python3 convert_bio.py --input_file data_dir --output_dir model_input_location
```
## Train model
```
python3 main.py --task youtube --model_dir youtube_model do_train = True do_eval = True
```
## Delete comments (Run evaluation)
```
python3 delete_comments.py --model_dir youtube_model --channel_id channelID --use_channel_id True
python3 delete_comments.py --model_dir youtube_model --video_id videoID --use_channel_id False
```
> Using the channel id will retrived the latest video in the specified channel.
## References
- [Huggingface Transformers](https://github.com/huggingface/transformers)
- [pytorch-crf](https://github.com/kmkurn/pytorch-crf)