Sea-Snell/Implicit-Language-Q-Learning
Official code from the paper "Offline RL for Natural Language Generation with Implicit Language Q Learning"
This project helps people who need to generate natural language responses that are not only grammatically correct but also strategically optimized for a specific goal, like producing engaging social media comments or generating helpful dialogue. It takes existing text data and a desired outcome (e.g., getting upvotes on Reddit, having a useful conversation) and trains a language model to produce new text that is highly likely to achieve that outcome. This is ideal for anyone working with conversational AI, content generation, or automated communication systems.
211 stars. No commits in the last 6 months.
Use this if you need to train a language model to generate text that optimizes for a specific, measurable real-world outcome, rather than just sounding human-like.
Not ideal if you're looking for a simple text generation tool without the need for sophisticated goal-oriented optimization or if you lack existing high-quality data and a reward mechanism to guide the model.
Stars
211
Forks
19
Language
Python
License
MIT
Category
Last pushed
Jul 31, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/Sea-Snell/Implicit-Language-Q-Learning"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
yaserkl/RLSeq2Seq
Deep Reinforcement Learning For Sequence to Sequence Models
kefirski/pytorch_RVAE
Recurrent Variational Autoencoder that generates sequential data implemented with pytorch
georgian-io/Multimodal-Toolkit
Multimodal model for text and tabular data with HuggingFace transformers as building block for text data
ctr4si/A-Hierarchical-Latent-Structure-for-Variational-Conversation-Modeling
PyTorch Implementation of "A Hierarchical Latent Structure for Variational Conversation...
nurpeiis/LeakGAN-PyTorch
A simple implementation of LeakGAN in PyTorch