Facebook AI has built and open-sourced BlenderBot, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators.
[...]
Today we’re releasing the complete model, code, and evaluation set-up, so that other AI researchers will be able to reproduce this work and continue to advance conversational AI research.
[...]
We used previously available public domain conversations that involved 1.5 billion training examples of extracted conversations.
It looks like there is a GitHub repo to check out and a Python command to run, but it requires a pretty hefty graphics card: https://parl.ai/projects/recipes/
It looks like there is a GitHub repo to check out and a Python command to run, but it requires a pretty hefty graphics card:
You'd think Facebook would have a few GPUs to spare to leave a running instance somewhere... More seriously, I think they probably want to tune it some more / gather feedback to avoid a disaster...
You'd think Facebook would have a few GPUs to spare to leave a running instance somewhere...
More seriously, I think they probably want to tune it some more / gather feedback to avoid a disaster like Microsoft Tay.
From the blog post:
[...]
[...]
Is it possible to try it somewhere ?
It looks like there is a GitHub repo to check out and a Python command to run, but it requires a pretty hefty graphics card:
https://parl.ai/projects/recipes/
You'd think Facebook would have a few GPUs to spare to leave a running instance somewhere...
More seriously, I think they probably want to tune it some more / gather feedback to avoid a disaster like Microsoft Tay.