ColossalAI
ploomber
Our great sponsors
ColossalAI | ploomber | |
---|---|---|
42 | 121 | |
37,836 | 3,369 | |
3.7% | 0.9% | |
9.7 | 7.8 | |
3 days ago | 17 days ago | |
Python | Python | |
Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
ColossalAI
- FLaNK AI-April 22, 2024
- Making large AI models cheaper, faster and more accessible
-
ColossalChat: An Open-Source Solution for Cloning ChatGPT with a RLHF Pipeline
> open-source a complete RLHF pipeline ... based on the LLaMA pre-trained model
I've gotten to where when I see "open source AI" I now know it's "well, except for $some_other_dependencies"
Anyway: https://scribe.rip/@yangyou_berkeley/colossalchat-an-open-so... and https://github.com/hpcaitech/ColossalAI#readme (Apache 2) can save you some medium.com heartache at least
-
Meet ColossalChat: An Open-Source AI Solution For Cloning ChatGPT With A Complete RLHF Pipeline
Quick Read: https://www.marktechpost.com/2023/04/01/meet-colossalchat-an-open-source-ai-solution-for-cloning-chatgpt-with-a-complete-rlhf-pipeline/ Github: https://github.com/hpcaitech/ColossalAI Examples: https://chat.colossalai.org/
-
A top AI researcher reportedly left Google for OpenAI after sharing concerns the company was training Bard on ChatGPT data
One of the current methods for training competing models is to have ChatGPT literally create prompt -> completion data sets. That's what was used for https://github.com/hpcaitech/ColossalAI. A model based off of the Llama weights released by facebook, then fine tuned on ChatGPT3.5 prompt + completions. So yes, there is a good chance that google is literally using ChatGPT in the training loop.
- Colossal-AI: open-source RLHF pipeline based on LLaMA pre-trained model
- ColossalChat
-
ColossalChat: An Open-Source Solution for Cloning ChatGPT with RLHF Pipeline
Here's the github from the article:
https://github.com/hpcaitech/ColossalAI
-
Open source solution replicates ChatGPT training process
The article talks about their RLHF implementation briefly. There’s details on their RLHF implementation here: https://github.com/hpcaitech/ColossalAI/blob/a619a190df71ea3...
-
how can I make my own chatGPT?
Here’s the project on GitHub: https://github.com/hpcaitech/ColossalAI
ploomber
-
Show HN: JupySQL – a SQL client for Jupyter (ipython-SQL successor)
- One-click sharing powered by Ploomber Cloud: https://ploomber.io
Documentation: https://jupysql.ploomber.io
Note that JupySQL is a fork of ipython-sql; which is no longer actively developed. Catherine, ipython-sql's creator, was kind enough to pass the project to us (check out ipython-sql's README).
We'd love to learn what you think and what features we can ship for JupySQL to be the best SQL client! Please let us know in the comments!
-
Runme – Interactive Runbooks Built with Markdown
For those who don't know, Jupyter has a bash kernel: https://github.com/takluyver/bash_kernel
And you can run Jupyter notebooks from the CLI with Ploomber: https://github.com/ploomber/ploomber
-
Rant: Jupyter notebooks are trash.
Develop notebook-based pipelines
-
Who needs MLflow when you have SQLite?
Fair point. MLflow has a lot of features to cover the end-to-end dev cycle. This SQLite tracker only covers the experiment tracking part.
We have another project to cover the orchestration/pipelines aspect: https://github.com/ploomber/ploomber and we have plans to work on the rest of features. For now, we're focusing on those two.
-
New to large SW projects in Python, best practices to organize code
I recommend taking a look at the ploomber open source. It helps you structure your code and parameterize it in a way that's easier to maintain and test. Our blog has lots of resources about it from testing your code to building a data science platform on AWS.
-
A three-part series on deploying a Data Science Platform on AWS
Developing end-to-end data science infrastructure can get complex. For example, many of us might have struggled to try to integrate AWS services and deal with configuration, permissions, etc. At Ploomber, we’ve worked with many companies in a wide range of industries, such as energy, entertainment, computational chemistry, and genomics, so we are constantly looking for simple solutions to get them started with Data Science in the cloud.
- Ploomber Cloud - Parametrizing and running notebooks in the cloud in parallel
-
Is Colab still the place to go?
If you like working locally with notebooks, you can run via the free tier of ploomber, that'll allow you to get the Ram/Compute you need for the bigger models as part of the free tier. Also, it has the historical executions so you don't need to remember what you executed an hour later!
-
Alternatives to nextflow?
It really depends on your use cases, I've seen a lot of those tools that lock you into a certain syntax, framework or weird language (for instance Groovy). If you'd like to use core python or Jupyter notebooks I'd recommend Ploomber, the community support is really strong, there's an emphasis on observability and you can deploy it on any executor like Slurm, AWS Batch or Airflow. In addition, there's a free managed compute (cloud edition) where you can run certain bioinformatics flows like Alphafold or Cripresso2
-
Saving log files
That's what we do for lineage with https://ploomber.io/
What are some alternatives?
DeepSpeed - DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
Kedro - Kedro is a toolbox for production-ready data science. It uses software engineering best practices to help you create data engineering and data science pipelines that are reproducible, maintainable, and modular.
Megatron-LM - Ongoing research training transformer models at scale
papermill - 📚 Parameterize, execute, and analyze notebooks
determined - Determined is an open-source machine learning platform that simplifies distributed training, hyperparameter tuning, experiment tracking, and resource management. Works with PyTorch and TensorFlow.
dagster - An orchestration platform for the development, production, and observation of data assets.
fairscale - PyTorch extensions for high performance and large scale training.
dvc - 🦉 ML Experiments and Data Management with Git
DeepFaceLive - Real-time face swap for PC streaming or video calls
argo - Workflow Engine for Kubernetes
PaddlePaddle - PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice (『飞桨』核心框架,深度学习&机器学习高性能单机、分布式训练和跨平台部署)
MLflow - Open source platform for the machine learning lifecycle