metadata

license: bsd-3-clause
datasets:
  - daily_dialog
language:
  - en
tags:
  - chat
  - dialog
  - natural language
  - codegen
widget:
  - text: |-
      Bot responds in friendly, longer messages.
      <Chat history>
      [User]: 'Hello!'
      [Bot]: 'Hi there! :D How are you?'
      [User]: 'Good, you?'
      [Bot]: '
    example_title: Greeting chat
co2_eq_emissions:
  emissions: 40
  source: https://mlco2.github.io/impact/#compute
  training_type: fine-tuning
  geographical_location: Oregon, USA
  hardware_used: 1x NVIDIA P100, Kaggle
library_name: transformers

DailyChat-350M

A finetuned version of Codegen-350M-nl on the 'daily_dialog' dataset. The idea of this model is to create one that is capable of holding a decent conversation.

Training Procedure

This was trained on Kaggle's servers using 1x NVIDIA P100. This model was trained for 1 epoch with learning rate 1e-2.

Biases & Limitations

This likely contains the same biases and limitations as the original model that it is based on, and additionally heavy biases from the dataset. It can generate offensive input when prompted, so user discretion is advised.

Intended Use

Dialog generation, chat agents.