All notebooks are beginner friendly! Add your dataset, click "Run All", and you'll get a 2x faster finetuned model which can be exported to GGUF, vLLM or uploaded to Hugging Face.
* Kaggle has 2x T4s, but we use 1. Due to overhead, 1x T4 is 5x faster.
SmolLM-1.7B-Instruct
SmolLM
Model Summary
SmolLM is a series of small language models available in three sizes: 135M, 360M, and 1.7B parameters.
These models are pre-trained on SmolLM-Corpus, a curated collection of high-quality educational and synthetic data designed for training LLMs. For further details, we refer to our blogpost.
To build SmolLM-Instruct, we finetuned the base models on publicly available datasets.
Changelog
Release
Description
v0.1
Initial release of SmolLM-Instruct. We finetune on the permissive subset of the WebInstructSub dataset, combined with StarCoder2-Self-OSS-Instruct. Then, we perform DPO (Direct Preference Optimization) for one epoch on HelpSteer for the 135M and 1.7B models, and argilla/dpo-mix-7k for the 360M model.
v0.2 models are better at staying on topic and responding appropriately to standard prompts, such as greetings and questions about their role as AI assistants. SmolLM-360M-Instruct (v0.2) has a 63.3% win rate over SmolLM-360M-Instruct (v0.1) on AlpacaEval. You can find the details here.
You can load v0.1 checkpoint by specifying revision="v0.1" in the transformers code:
model = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/SmolLM-1.7B-Instruct", revision="v0.1")
We noticed that 4bit quantization degrades the quality of the 135M and 360M, so we use q016 for MLC and ONNX/Transformers.js checkpoints for the WebGPU demos. We also suggest using temperature 0.2 and top-p 0.9.
Transformers
pip install transformers
python
1# pip install transformers2from transformers import AutoModelForCausalLM, AutoTokenizer
3checkpoint ="HuggingFaceTB/SmolLM-1.7B-Instruct"45device ="cuda"# for GPU usage or "cpu" for CPU usage6tokenizer = AutoTokenizer.from_pretrained(checkpoint)7# for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")`8model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)910messages =[{"role":"user","content":"What is the capital of France."}]11input_text=tokenizer.apply_chat_template(messages, tokenize=False)12print(input_text)13inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)14outputs = model.generate(inputs, max_new_tokens=50, temperature=0.2, top_p=0.9, do_sample=True)15print(tokenizer.decode(outputs[0]))
Chat in TRL
You can also use the TRL CLI to chat with the model from the terminal:
bash
1pip install trl
2trl chat --model_name_or_path HuggingFaceTB/SmolLM-1.7B-Instruct --device cpu
Limitations
Additionally, the generated content may not always be factually accurate, logically consistent, or free from biases present in the training data, we invite users to leverage them as assistive tools rather than definitive sources of information. We find that they can handle general knowledge questions, creative writing and basic Python programming. But they are English only and may have difficulty with arithmetics, editing tasks and complex reasoning. For more details about the models' capabilities, please refer to our blog post.
Training parameters
We train the models using the alignment-handbook with the datasets mentioned in the changelog, using these parameters v0.2 (most of them are from Zephyr Gemma recipe):
1@misc{allal2024SmolLM,
2title={SmolLM - blazingly fast and remarkably powerful},
3author={Loubna Ben Allal and Anton Lozhkov and Elie Bakouch and Leandro von Werra and Thomas Wolf},
4year={2024},
5}
