Instructions to use tiiuae/Falcon-E-3B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiiuae/Falcon-E-3B-Instruct with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="tiiuae/Falcon-E-3B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tiiuae/Falcon-E-3B-Instruct")
model = AutoModelForCausalLM.from_pretrained("tiiuae/Falcon-E-3B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use tiiuae/Falcon-E-3B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiiuae/Falcon-E-3B-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiiuae/Falcon-E-3B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tiiuae/Falcon-E-3B-Instruct

SGLang

How to use tiiuae/Falcon-E-3B-Instruct 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-E-3B-Instruct" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiiuae/Falcon-E-3B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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-E-3B-Instruct" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiiuae/Falcon-E-3B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tiiuae/Falcon-E-3B-Instruct with Docker Model Runner:
```
docker model run hf.co/tiiuae/Falcon-E-3B-Instruct
```

Does it work on CPU?

by LLMToaster - opened May 16, 2025

Discussion

LLMToaster

May 16, 2025

Once again thank you for creating such a great model. Does it work on CPU or does it require modern architecture (ampere or later), how about intel or AMD GPUs?

Also, same question about 1.58bit, Falcon3 10B 1.58bit model will that work on CPU?

ybelkada

Technology Innovation Institute org May 16, 2025

Hi @LLMToaster
Thanks for your question !
Yes, you can either run it with HF transformers on CPU (though being not optimized), or use bitnet.cpp (recommended)
For bitnet.cpp, you can go ahead and use: https://huggingface.co/tiiuae/Falcon-E-1B-Instruct-GGUF together with: https://github.com/microsoft/BitNet/pull/268 (right now the PR has not been merged)

ybelkada

Technology Innovation Institute org May 16, 2025

For the 10B model, you can use: https://huggingface.co/tiiuae/Falcon3-10B-Instruct-1.58bit and no need to use https://github.com/microsoft/BitNet/pull/268 - bitnet main already supports Falcon3 models

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment