Oleg Polakow

In the next version, I will add `pass_pd` argument to pass pandas objects instead of numpy arrays.

There is no need for 4 different boolean arrays or an array with integer type, just use other arguments such as direction to guide your signals.

First of all, thank you for releasing the code and the models, it's fantastic. Based on the current DialoGPT implementation, I adapted [run_generation.py](https://github.com/huggingface/transformers/blob/master/examples/run_generation.py) from Hugging Face to perform decoding and...

You need to adapt [interactive_bot.py](https://github.com/polakowo/gpt2bot/blob/master/gpt2bot/interactive_bot.py) by creating a route around the following lines https://github.com/polakowo/gpt2bot/blob/a091b7ec7b350819ca5cf3efa5675ed420890fc8/gpt2bot/interactive_bot.py#L35-L66 Having query `prompt` it will generate a bot message `bot_message` that you can send back to...

Looks good. You would also need some sort of session management and ideally a route for updating the parameters of the generator such as temperature, no need to restart the...

The best you can do is to fine tune it on your custom dataset, even a few texts (

Yes, I'm going to update the package to the new transformers version soon.

I released a new version that accepts any text generation model, including large dialogpt (use `model = microsoft/DialoGPT-large` in chatbot.cfg).

@lucmichalski looks great! I had a Dockerfile previously but it downloaded the same model every time I deployed the container, which made it too expensive to play with. Maybe I...

Hi, it stores messages per chat (see `chat_data` in [documentation](https://github.com/python-telegram-bot/python-telegram-bot/wiki/Storing-bot%2C-user-and-chat-related-data)), which makes sense since bot must be able to communicate to multiple users in a group. If you want to...