bagelnet/df67eac8-f2e7-4f5e-a0de-b2ae24fb8579

SmolLM2

Table of Contents

1. Model Summary
2. Evaluation
3. Limitations
4. Training
5. License
6. Citation

Model Summary

SmolLM2 is a family of compact language models available in three size: 135M, 360M, and 1.7B parameters. They are capable of solving a wide range of tasks while being lightweight enough to run on-device.

The 1.7B variant demonstrates significant advances over its predecessor SmolLM1-1.7B, particularly in instruction following, knowledge, reasoning, and mathematics. It was trained on 11 trillion tokens using a diverse dataset combination: FineWeb-Edu, DCLM, The Stack, along with new mathematics and coding datasets that we curated and will release soon. We developed the instruct version through supervised fine-tuning (SFT) using a combination of public datasets and our own curated datasets. We then applied Direct Preference Optimization (DPO) using UltraFeedback.

The instruct model additionally supports tasks such as text rewriting, summarization and function calling thanks to datasets developed by Argilla such as Synth-APIGen-v0.1.

How to use

pip install transformers

Running the model on CPU/GPU/multi GPU

• Using full precision

# pip install transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
checkpoint = "HuggingFaceTB/SmolLM2-1.7B"
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("Gravity is", return_tensors="pt").to(device)
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))

• Using torch.bfloat16

# pip install accelerate
# for fp16 use `torch_dtype=torch.float16` instead
model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto", torch_dtype=torch.bfloat16)
inputs = tokenizer.encode("Gravity is", return_tensors="pt").to("cuda")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))

>>> print(f"Memory footprint: {model.get_memory_footprint() / 1e6:.2f} MB")
Memory footprint: 3422.76 MB

Evaluation

In this section, we report the evaluation results of SmolLM2. All evaluations are zero-shot unless stated otherwise, and we use lighteval to run them.

Base Pre-Trained Model

Metric	SmolLM2-1.7B	Llama-1B	Qwen2.5-1.5B	SmolLM1-1.7B
HellaSwag	68.7	61.2	66.4	62.9
ARC (Average)	60.5	49.2	58.5	59.9
PIQA	77.6	74.8	76.1	76.0
MMLU-Pro (MCF)	19.4	11.7	13.7	10.8
CommonsenseQA	43.6	41.2	34.1	38.0
TriviaQA	36.7	28.1	20.9	22.5
Winogrande	59.4	57.8	59.3	54.7
OpenBookQA	42.2	38.4	40.0	42.4
GSM8K (5-shot)	31.0	7.2	61.3	5.5

Instruction Model

Metric	SmolLM2-1.7B-Instruct	Llama-1B-Instruct	Qwen2.5-1.5B-Instruct	SmolLM1-1.7B-Instruct
IFEval (Average prompt/inst)	56.7	53.5	47.4	23.1
MT-Bench	6.13	5.48	6.52	4.33
OpenRewrite-Eval (micro_avg RougeL)	44.9	39.2	46.9	NaN
HellaSwag	66.1	56.1	60.9	55.5
ARC (Average)	51.7	41.6	46.2	43.7
PIQA	74.4	72.3	73.2	71.6
MMLU-Pro (MCF)	19.3	12.7	24.2	11.7
BBH (3-shot)	32.2	27.6	35.3	25.7
GSM8K (5-shot)	48.2	26.8	42.8	4.62

Limitations

SmolLM2 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.

Training

Model

• Architecture: Transformer decoder
• Pretraining tokens: 11T
• Precision: bfloat16

Hardware

• GPUs: 256 H100

Software

• Training Framework: nanotron

License

Apache 2.0

Citation

@misc{allal2024SmolLM2,
      title={SmolLM2 - with great data, comes great performance}, 
      author={Loubna Ben Allal and Anton Lozhkov and Elie Bakouch and Gabriel Martín Blázquez and Lewis Tunstall and Agustín Piqueres and Andres Marafioti and Cyril Zakka and Leandro von Werra and Thomas Wolf},
      year={2024},
}