Instructions to use tiiuae/falcon-180B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use tiiuae/falcon-180B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="tiiuae/falcon-180B")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("tiiuae/falcon-180B") model = AutoModelForCausalLM.from_pretrained("tiiuae/falcon-180B") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use tiiuae/falcon-180B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "tiiuae/falcon-180B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "tiiuae/falcon-180B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/tiiuae/falcon-180B
- SGLang
How to use tiiuae/falcon-180B with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "tiiuae/falcon-180B" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "tiiuae/falcon-180B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "tiiuae/falcon-180B" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "tiiuae/falcon-180B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use tiiuae/falcon-180B with Docker Model Runner:
docker model run hf.co/tiiuae/falcon-180B
The email from the license seems unreachable
Hi!
/* first of all, thank you a lot for releasing falcon-180B. This is a great step towards open science and democratizing LLMs - and a great precedent for future releases */
We'd love to add falconllm-180B to Petals (a non-commercial research project) - https://github.com/bigscience-workshop/petals .
This would allow people to run experiments with falcon-180B directly from their laptops or free colab instances.
We're concerned that our project technically contains a library-specific API (see readme). The purpose of this API is to let users run the distributed model from python in a way that is compatible with transformers (see code example). The API is open-source and non-commercial, but it could theoretically fall under sections 9.1-9.4 of the license:
9.1: Hosting Use means any use of the Work or a Derivative Work to offer shared instances ... in an inference or finetuning API form
The license suggests that we send an email to Falconllm.partnerships@tii.ae and ask for permission.
Unfortunately, the emails don't seem to get through. We get the following error message from all major email services:
Q1: can we use falcon-180B in Petals?
Q2: can you please check that the email is set up and reachable?
