metadata
library_name: transformers
tags:
- falcon3
base_model: tiiuae/Falcon3-10B-Base
license: other
license_name: falcon-llm-license
license_link: https://falconllm.tii.ae/falcon-terms-and-conditions.html
Falcon3-10B-Instruct
Falcon3 family of Open Foundation Models is a set of pretrained and instruct LLMs ranging from 1B to 10B parameters.
This repository contains the Falcon3-10B-Instruct. It achieves state-of-the-art results (at the time of release) on reasoning, language understanding, instruction following, code and mathematics tasks. Falcon3-10B-Instruct supports 4 languages (English, French, Spanish, Portuguese) and a context length of up to 32K.
Model Details
- Architecture
- Transformer-based causal decoder-only architecture
- 40 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
- Depth up-scaled from Falcon3-7B-Base with 2 Teratokens of datasets comprising of web, code, STEM, high quality and mutlilingual data using 1024 H100 GPU chips
- Posttrained on 1.2 million samples of STEM, conversational, 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/Falcon3-10B-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 | Yi-1.5-9B-Chat | Mistral-Nemo-Base-2407 (12B) | Falcon3-10B-Instruct |
|---|---|---|---|---|
| General | MMLU (5-shot) | 70 | 65.9 | 71.6 |
| MMLU-PRO (5-shot) | 39.6 | 32.7 | 44 | |
| IFEval | 57.6 | 63.4 | 78 | |
| Math | GSM8K (5-shot) | 76.6 | 73.8 | 83.1 |
| GSM8K (8-shot, COT) | 78.5 | 73.6 | 81.3 | |
| MATH Lvl-5 (4-shot) | 8.8 | 0.4 | 22.1 | |
| Reasoning | Arc Challenge (25-shot) | 51.9 | 61.6 | 64.5 |
| GPQA (0-shot) | 35.4 | 33.2 | 33.5 | |
| GPQA (0-shot, COT) | 16 | 12.7 | 32.6 | |
| MUSR (0-shot) | 41.9 | 38.1 | 41.1 | |
| BBH (3-shot) | 49.2 | 43.6 | 58.4 | |
| CommonSense Understanding | PIQA (0-shot) | 76.4 | 78.2 | 78.4 |
| SciQ (0-shot) | 61.7 | 76.4 | 90.4 | |
| Winogrande (0-shot) | - | - | 71.3 | |
| OpenbookQA (0-shot) | 43.2 | 47.4 | 48.2 | |
| Instructions following | MT-Bench (avg) | 8.28 | 8.6 | 8.17 |
| Alpaca (WC) | 25.81 | 45.44 | 24.7 | |
| Tool use | BFCL AST (avg) | 48.4 | 74.2 | 86.3 |
| Code | EvalPlus (0-shot) (avg) | 69.4 | 58.9 | 74.7 |
| Multipl-E (0-shot) (avg) | - | 34.5 | 45.8 |
Technical Report
Coming soon....
Citation
If Falcon3 family were helpful in 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}
}