Using datasets from GPT-2 to display randomized text between Webtext and generated texts.
- Choose a difficulty and guess which text was written by a human and which was made by a computer.
- Have points for the number of correct guesses in a row
- Leaderboard
Thanks to OpenAI for releasing a subset of their data.
Please forward any feedback to jun.a.kagaya@gmail.com
Buy me a coffee! Patreon
$ conda create -n whichtextisreal_env $ conda activate whichtextisreal_env $ conda deactivate gcloud domains list-user-verified https://storage.googleapis.com/gpt-2/output-dataset/v1/webtext.train.jsonl gcloud auth application-default logingcloud config set project <isthistextreal gcloud initgcloud app deploy app.yaml pip install -t lib -r requirements.txt —upgrade dev_appserver.py app.yaml brew install python@2 export PATH="/usr/local/opt/python@2/libexec/bin:$PATH" export CLOUDSDK_PYTHON="/usr/local/opt/python@2/libexec/bin:$PATH" export CLOUDSDK_PYTHON="/usr/local/bin/python" cd ~/desktop/"openai gpt2"/whichtextisreal.com $ pip install -t lib -r requirements.txt --upgrade gcloud app browse gcloud app logs tail -s default