Run 100B+ language models at home, BitTorrent-style.
Fine-tuning and inference up to 10x faster than offloading

Generate text using distributed 176B-parameter BLOOM or BLOOMZ and fine-tune them for your own tasks:

from petals import DistributedBloomForCausalLM

model = DistributedBloomForCausalLM.from_pretrained("bigscience/bloom-petals", tuning_mode="ptune", pre_seq_len=16)
# Embeddings & prompts are on your device, BLOOM blocks are distributed across the Internet

inputs = tokenizer("A cat sat", return_tensors="pt")["input_ids"]
outputs = model.generate(inputs, max_new_tokens=5)
print(tokenizer.decode(outputs[0]))  # A cat sat on a mat...

# Fine-tuning (updates only prompts or adapters hosted locally)
optimizer = torch.optim.AdamW(model.parameters())
for input_ids, labels in data_loader:
    outputs = model.forward(input_ids)
    loss = cross_entropy(outputs.logits, labels)
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()

🚀  Try now in Colab

🔏 Your data will be processed by other people in the public swarm. Learn more about privacy here. For sensitive data, you can set up a private swarm among people you trust.

Connect your GPU and increase Petals capacity

Run this in an Anaconda env (requires Linux and Python 3.7+):

conda install pytorch pytorch-cuda=11.7 -c pytorch -c nvidia
pip install -U petals
python -m petals.cli.run_server bigscience/bloom-petals

Or use our Docker image (works on Linux, macOS, and Windows with WSL2):

sudo docker run -p 31330:31330 --ipc host --gpus all --volume petals-cache:/cache --rm \
    learningathome/petals:main python -m petals.cli.run_server bigscience/bloom-petals --port 31330

📚 See FAQ to learn how to configure the server to use multiple GPUs, address common issues, etc.

You can also host BLOOMZ, a version of BLOOM fine-tuned to follow human instructions in the zero-shot regime — just replace bloom-petals with bloomz-petals.

🔒 Hosting a server does not allow others to run custom code on your computer. Learn more about security here.

💬 If you have any issues or feedback, let us know on our Discord server!

Check out tutorials, examples, and more

Basic tutorials:

• Getting started: tutorial
• Prompt-tune BLOOM to create a personified chatbot: tutorial
• Prompt-tune BLOOM for text semantic classification: tutorial

Useful tools and advanced guides:

• Chatbot web app (connects to Petals via an HTTP endpoint): source code
• Monitor for the public swarm: source code
• Launch your own swarm: guide
• Run a custom foundation model: guide

Learning more:

• Frequently asked questions: FAQ
• In-depth system description: paper

📋 If you build an app running BLOOM with Petals, make sure it follows the BLOOM's terms of use.

How does it work?

• Petals runs large language models like BLOOM-176B collaboratively — you load a small part of the model, then team up with people serving the other parts to run inference or fine-tuning.
• Single-batch inference runs at ≈ 1 sec per step (token) — up to 10x faster than offloading, enough for chatbots and other interactive apps. Parallel inference reaches hundreds of tokens/sec.
• Beyond classic language model APIs — you can employ any fine-tuning and sampling methods, execute custom paths through the model, or see its hidden states. You get the comforts of an API with the flexibility of PyTorch.

📚  See FAQ            📜  Read paper

Installation

Here's how to install Petals with Anaconda on Linux:

conda install pytorch pytorch-cuda=11.7 -c pytorch -c nvidia
pip install -U petals

If you don't use Anaconda, you can install PyTorch in any other way. If you want to run models with 8-bit weights, please install PyTorch with CUDA 11.x or newer for compatility with bitsandbytes.

See the instructions for macOS and Windows, the full requirements, and troubleshooting advice in our FAQ.

⏱️ Benchmarks

Network		Single-batch inference (steps/s)		Parallel forward (tokens/s)
Bandwidth	Round-trip latency	Sequence length		Batch size
		128	2048	1	64
Offloading, max. possible speed on 1x A100 1
256 Gbit/s		0.18	0.18	2.7	170.3
128 Gbit/s		0.09	0.09	2.4	152.8
Petals on 14 heterogeneous servers across Europe and North America 2
Real world		0.83	0.79	32.6	179.4
Petals on 3 servers, with one A100 each 3
1 Gbit/s	< 5 ms	1.71	1.54	70.0	253.6
100 Mbit/s	< 5 ms	1.66	1.49	56.4	182.0
100 Mbit/s	100 ms	1.23	1.11	19.7	112.2

¹ An upper bound for offloading performance. We base our offloading numbers on the best possible hardware setup for offloading: CPU RAM offloading via PCIe 4.0 with 16 PCIe lanes per GPU and PCIe switches for pairs of GPUs. We assume zero latency for the upper bound estimation. In 8-bit, the model uses 1 GB of memory per billion parameters. PCIe 4.0 with 16 lanes has a throughput of 256 Gbit/s, so offloading 176B parameters takes 5.5 seconds. The throughput is twice as slow (128 Gbit/s) if we have two GPUs behind the same PCIe switch.

² A real-world distributed setting with 14 servers holding 2× RTX 3060, 4× 2080Ti, 2× 3090, 2× A4000, and 4× A5000 GPUs. These are personal servers and servers from university labs, spread across Europe and North America and connected to the Internet at speeds of 100–1000 Mbit/s. 4 servers operate from under firewalls.

³ An optimistic setup that requires least communication. The client nodes have 8 CPU cores and no GPU.

We provide more evaluations and discuss these results in more detail in Section 3.3 of our paper.

🛠️ Contributing

Please see our FAQ on contributing.

📜 Citation

Alexander Borzunov, Dmitry Baranchuk, Tim Dettmers, Max Ryabinin, Younes Belkada, Artem Chumachenko, Pavel Samygin, and Colin Raffel. Petals: Collaborative Inference and Fine-tuning of Large Models. arXiv preprint arXiv:2209.01188, 2022.

@article{borzunov2022petals,
  title = {Petals: Collaborative Inference and Fine-tuning of Large Models},
  author = {Borzunov, Alexander and Baranchuk, Dmitry and Dettmers, Tim and Ryabinin, Max and Belkada, Younes and Chumachenko, Artem and Samygin, Pavel and Raffel, Colin},
  journal = {arXiv preprint arXiv:2209.01188},
  year = {2022},
  url = {https://arxiv.org/abs/2209.01188}
}

---

This project is a part of the BigScience research workshop.