metadata
language:
- en
- fr
- es
- pt
tags:
- falcon3
base_model: tiiuae/Falcon3-7B-Base
license: other
license_name: falcon-llm-license
license_link: https://falconllm.tii.ae/falcon-terms-and-conditions.html
Falcon3-7B-Instruct
Falcon3 family of Open Foundation Models is a set of pretrained and instruct LLMs ranging from 1B to 10B.
This repository contains the Falcon3-7B-Instruct. It achieves state of art results (at release's time) on reasoning, language understanding, instruction following, code and mathematics tasks. Falcon3-7B-Instruct supports 4 languages (english, french, spanish, portuguese) and a context length up to 32K.
Model Details
- Architecture
- Transformer based causal decoder only architecture
- 28 decoder blocks
- Grouped query attention (GQA) for faster inference: 12 query heads and 4 key value heads
- Wider head dimension: 256
- High RoPE value to support long context understanding: 1000042
- Uses SwiGLU and RMSNorm
- 32K context length
- 131K vocab size
- Pretrained on 14 Teratokens of datasets comprising of web, code, STEM, high quality and mutlilingual data using 2048 H100 GPU chips
- Postrained on 1.2 million samples of STEM, conversations, code, safety and function call data
- Supports EN, FR, ES, PT
- Developed by Technology Innovation Institute
- License: TII Falcon-LLM License 2.0
- Model Release Date: December 2024
Getting started
Click to expand
from transformers import AutoTokenizer, AutoModelForCausalLM
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "tiiuae/Falcon3-7B-Instruct"
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype="auto",
device_map="auto"]
)
tokenizer = AutoTokenizer.from_pretrained(model_name)
prompt = "How many hours in one day?"
messages = [
{"role": "system", "content": "You are a helpful friendly assistant Falcon3 from TII, try to follow instructions as much as possible."},
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
generated_ids = model.generate(
**model_inputs,
max_new_tokens=1024
)
generated_ids = [
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(response)
Benchmarks
We report in the following table our internal pipeline benchmarks:
| Category | Benchmark | Llama-3.2-1B | Qwen2.5-1.5B | SmolLM2-1.7B | Falcon3-1B-Instruct |
|---|---|---|---|---|---|
| General | MMLU (5-shot) | 23.4 | 58.4 | 48.4 | 43.9 |
| MMLU-PRO (5-shot) | 11.3 | 21.3 | 17.2 | 18.6 | |
| IFEval | 55.8 | 44.4 | 53.0 | 54.4 | |
| Math | GSM8K (5-shot) | 37.4 | 57.2 | 43.4 | 38.6 |
| GSM8K (8-shot, COT) | 35.6 | 62.2 | 47.2 | 41.8 | |
| MATH Lvl-5 (4-shot) | 3.9 | 0.2 | 0.1 | 1.0 | |
| Reasoning | Arc Challenge (25-shot) | 34.1 | 47.0 | 47.6 | 45.9 |
| GPQA (0-shot) | 25.3 | 29.6 | 28.7 | 26.5 | |
| GPQA (0-shot, COT) | 13.2 | 9.2 | 16.0 | 21.3 | |
| MUSR (0-shot) | 32.4 | 36.8 | 33.0 | 40.7 | |
| BBH (3-shot) | 30.3 | 38.5 | 33.1 | 35.1 | |
| BBH (3-shot, COT) | 0.0 | 20.3 | 0.8 | 30.5 | |
| CommonSense Understanding | PIQA (0-shot) | 72.1 | 73.2 | 74.4 | 72.0 |
| SciQ (0-shot) | 61.8 | 69.5 | 71.4 | 86.8 | |
| Winogrande (0-shot) | - | - | - | 60.2 | |
| OpenbookQA (0-shot) | 40.2 | 40.4 | 42.8 | 40.0 | |
| MT-Bench (avg) | 5.4 | 7.1 | 6.1 | 5.5 | |
| Instructions following | Alapaca (WC) | 8.6 | 8.6 | 5.4 | 6.1 |
Technical Report
Coming soon....
Citation
If Falcon3 family were helpful to your work, feel free to give us a cite.
@misc{Falcon3,
title = {The Falcon 3 family of Open Models},
author = {TII Team},
month = {December},
year = {2024}
}