Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/huangwl18/language-planner
Official Code for "Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents"
https://github.com/huangwl18/language-planner
artificial-intelligence codex deep-learning embodied-ai foundation-models gpt-3 in-context-learning knowledge-extraction language-model planning transformers
Last synced: 5 days ago
JSON representation
Official Code for "Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents"
- Host: GitHub
- URL: https://github.com/huangwl18/language-planner
- Owner: huangwl18
- License: mit
- Created: 2022-01-03T16:11:22.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2022-05-16T00:24:12.000Z (over 2 years ago)
- Last Synced: 2024-08-02T18:43:18.294Z (3 months ago)
- Topics: artificial-intelligence, codex, deep-learning, embodied-ai, foundation-models, gpt-3, in-context-learning, knowledge-extraction, language-model, planning, transformers
- Language: Jupyter Notebook
- Homepage: https://huangwl18.github.io/language-planner/
- Size: 20.3 MB
- Stars: 235
- Watchers: 4
- Forks: 31
- Open Issues: 5
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
[](https://colab.research.google.com/github/huangwl18/language-planner/blob/main/src/demo.ipynb)
## Language Models as Zero-Shot Planners:
Extracting Actionable Knowledge for Embodied Agents
#### [[Project Page]](https://huangwl18.github.io/language-planner/) [[Paper]](https://arxiv.org/pdf/2201.07207.pdf) [[Video]](https://www.youtube.com/watch?v=CkyugWI3_fc)
[Wenlong Huang](https://wenlong.page)1, [Pieter Abbeel](http://people.eecs.berkeley.edu/~pabbeel/)1, [Deepak Pathak](https://www.cs.cmu.edu/~dpathak/)\*2, [Igor Mordatch](https://scholar.google.com/citations?user=Vzr1RukAAAAJ&hl=en)\*3 (*equal advising)
1University of California, Berkeley, 2Carnegie Mellon University, 3Google Brain
This is the official demo code for our [Language Models as Zero-Shot Planners](https://huangwl18.github.io/language-planner/) paper. The code demonstrates how Large Language Models, such as GPT-3 and Codex, can generate action plans for complex human activities (e.g. "make breakfast"), even without any further training. The code can be used with any available language models from [OpenAI API](https://openai.com/api/) and [Huggingface Transformers](https://huggingface.co/docs/transformers/index) with a common interface.
If you find this work useful in your research, please cite using the following BibTeX:
```bibtex
@article{huang2022language,
title={Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents},
author={Huang, Wenlong and Abbeel, Pieter and Pathak, Deepak and Mordatch, Igor},
journal={arXiv preprint arXiv:2201.07207},
year={2022}
}
```
## Local Setup or [Open in Colab](https://colab.research.google.com/github/huangwl18/language-planner/blob/main/src/demo.ipynb)
### Requirements
- Python=3.6.13
- CUDA=11.3
### Setup Instructions
```Shell
git clone https://github.com/huangwl18/language-planner.git
cd language-planner/
conda create --name language-planner-env python=3.6.13
conda activate language-planner-env
pip install --upgrade pip
pip install -r requirements.txt
```
## Running Code
See [`demo.ipynb`](https://github.com/huangwl18/language-planner/blob/main/src/demo.ipynb) (or [](https://colab.research.google.com/github/huangwl18/language-planner/blob/main/src/demo.ipynb)) for a complete walk-through of our method. Feel free to experiment with any household tasks that you come up with (or any tasks beyond household domain if you provide necessary actions in [`available_actions.json`](https://github.com/huangwl18/language-planner/blob/main/src/available_actions.ipynb))!
**Note:**
- It is observed that best results can be obtained with larger language models. If you cannot run [Huggingface Transformers](https://huggingface.co/models?pipeline_tag=text-generation&sort=downloads) models locally or on Google Colab due to memory constraint, it is recommended to register an [OpenAI API](https://openai.com/api/) account and use GPT-3 or Codex (As of 01/2022, $18 free credits are awarded to new accounts and Codex series are free after [admitted from the waitlist](https://share.hsforms.com/1GzaACuXwSsmLKPfmphF_1w4sk30?)).
- Due to language models' high sensitivity to sampling hyperparameters, you may need to tune sampling hyperparameters for different models to obtain the best results.
- The code uses the list of available actions supported in [VirtualHome 1.0](https://github.com/xavierpuigf/virtualhome/tree/v1.0.0)'s [Evolving Graph Simulator](https://github.com/xavierpuigf/virtualhome/tree/v1.0.0/simulation). The available actions are stored in [`available_actions.json`](https://github.com/huangwl18/language-planner/blob/main/src/available_actions.json). The actions should support a large variety of household tasks. However, you may modify or replace this file if you're interested in a different set of actions or a different domain of tasks (beyond household domain).
- A subset of the [manually-annotated examples](http://virtual-home.org/release/programs/programs_processed_precond_nograb_morepreconds.zip) originally collected by the [VirtualHome paper](https://arxiv.org/pdf/1806.07011.pdf) is used as available examples in the prompt. They are transformed to natural language format and stored in [`available_examples.json`](https://github.com/huangwl18/language-planner/blob/main/src/available_examples.json). Feel free to change this file for a different set of available examples.