SBWBackend

FastAPI server to make requests to finetuned model based on ru_transformers. Model is finetuned on 750 mb of My Little Pony fanfics using v100 GPU which you can access here.
Text corpus is from ponyfiction.
DockerHub with gpu support is here and only cpu-version is here.
Use /generate POST method with payload like that:

{"text": "Твайлайт Спаркл удивлённо ", "lenght": 40, "temperature": 1}

To generate. Params of request are:

Input:
    text (str): input text
    length (int): lenght of generation in tokens
    temperature (float): "creativness" of generation
Returns:
    str: generated text