---
library_name: transformers
license: apache-2.0
language:
- en
---


# SmolLM


##  Table of Contents

1. [Model Summary](##model-summary)
2. [Limitations](##limitations)
3. [Training](##training)
4. [License](##license)
5. [Citation](##citation)

## Model Summary

SmolLM is a series of state-of-the-art small language models available in three sizes: 135M, 360M, and 1.7B parameters. These models are built on Cosmo-Corpus, a meticulously curated high-quality training dataset. Cosmo-Corpus includes Cosmopedia v2 (28B tokens of synthetic textbooks and stories generated by Mixtral), Python-Edu (4B tokens of educational Python samples from The Stack), and FineWeb-Edu (220B tokens of deduplicated educational web samples from FineWeb). SmolLM models have shown promising results when compared to other models in their size categories across various benchmarks testing common sense reasoning and world knowledge. For detailed information on training, benchmarks and performance, please refer to our full blog post ADD LINK WHEN PUBLISH.


### Generation
First, make sure to install `transformers` from source:
```bash
pip install git+https://github.com/huggingface/transformers.git
```

#### Running the model on CPU/GPU/multi GPU
* _Using full precision_
```python
# pip install git+https://github.com/huggingface/transformers.git # TODO: merge PR to main
from transformers import AutoModelForCausalLM, AutoTokenizer
checkpoint = "HuggingFaceTB/SmolLM-135M"
device = "cuda" # for GPU usage or "cpu" for CPU usage
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
# for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")`
model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)
inputs = tokenizer.encode("def print_hello_world():", return_tensors="pt").to(device)
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
```
```bash
>>> print(f"Memory footprint: {model.get_memory_footprint() / 1e6:.2f} MB")
Memory footprint: 12624.81 MB
```
* _Using `torch.bfloat16`_
```python
# pip install accelerate
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
checkpoint = "HuggingFaceTB/SmolLM-135M"
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
# for fp16 use `torch_dtype=torch.float16` instead
model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto", torch_dtype=torch.bfloat16)
inputs = tokenizer.encode("def print_hello_world():", return_tensors="pt").to("cuda")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
```
```bash
>>> print(f"Memory footprint: {model.get_memory_footprint() / 1e6:.2f} MB")
Memory footprint: 269.03 MB
```

#### Quantized Versions through `bitsandbytes`
* _Using 8-bit precision (int8)_

```python
# pip install bitsandbytes accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
# to use 4bit use `load_in_4bit=True` instead
quantization_config = BitsAndBytesConfig(load_in_8bit=True)
checkpoint = "HuggingFaceTB/SmolLM-135M"
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
model = AutoModelForCausalLM.from_pretrained(checkpoint, quantization_config=quantization_config)
inputs = tokenizer.encode("def print_hello_world():", return_tensors="pt").to("cuda")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
```
```bash
>>> print(f"Memory footprint: {model.get_memory_footprint() / 1e6:.2f} MB")
# load_in_8bit
Memory footprint: 162.87 MB
# load_in_4bit
>>> print(f"Memory footprint: {model.get_memory_footprint() / 1e6:.2f} MB")
Memory footprint: 109.78 MB
```

# Limitations

While SmolLM models have been trained on a diverse dataset including educational content and synthetic texts, they have limitations. The models primarily understand and generate content in English. They can produce text on a variety of topics, but the generated content may not always be factually accurate, logically consistent, or free from biases present in the training data. These models should be used as assistive tools rather than definitive sources of information. Users should always verify important information and critically evaluate any generated content. For a more comprehensive discussion of the models' capabilities and limitations, please refer to our full blog post.

# Training

## Model

- **Architecture:** For architecture detail, see the blog post
- **Pretraining steps:** 600k
- **Pretraining tokens:** 600B
- **Precision:** bfloat16

## Hardware

- **GPUs:** 64 H100

## Software

- **Training Framework:** [Nanotron](https://github.com/huggingface/nanotron/tree/main)

# License

[Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0)

# Citation
TO MODIFY
```bash
@misc{lozhkov2024starcoder,
      title={StarCoder 2 and The Stack v2: The Next Generation}, 
      author={Anton Lozhkov and Raymond Li and Loubna Ben Allal and Federico Cassano and Joel Lamy-Poirier and Nouamane Tazi and Ao Tang and Dmytro Pykhtar and Jiawei Liu and Yuxiang Wei and Tianyang Liu and Max Tian and Denis Kocetkov and Arthur Zucker and Younes Belkada and Zijian Wang and Qian Liu and Dmitry Abulkhanov and Indraneil Paul and Zhuang Li and Wen-Ding Li and Megan Risdal and Jia Li and Jian Zhu and Terry Yue Zhuo and Evgenii Zheltonozhskii and Nii Osae Osae Dade and Wenhao Yu and Lucas Krauß and Naman Jain and Yixuan Su and Xuanli He and Manan Dey and Edoardo Abati and Yekun Chai and Niklas Muennighoff and Xiangru Tang and Muhtasham Oblokulov and Christopher Akiki and Marc Marone and Chenghao Mou and Mayank Mishra and Alex Gu and Binyuan Hui and Tri Dao and Armel Zebaze and Olivier Dehaene and Nicolas Patry and Canwen Xu and Julian McAuley and Han Hu and Torsten Scholak and Sebastien Paquet and Jennifer Robinson and Carolyn Jane Anderson and Nicolas Chapados and Mostofa Patwary and Nima Tajbakhsh and Yacine Jernite and Carlos Muñoz Ferrandis and Lingming Zhang and Sean Hughes and Thomas Wolf and Arjun Guha and Leandro von Werra and Harm de Vries},
      year={2024},
      eprint={2402.19173},
      archivePrefix={arXiv},
      primaryClass={cs.SE}
}
```