Instructions to use colinglab/CLASS_IT-140M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use colinglab/CLASS_IT-140M with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="colinglab/CLASS_IT-140M")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("colinglab/CLASS_IT-140M") model = AutoModelForCausalLM.from_pretrained("colinglab/CLASS_IT-140M") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use colinglab/CLASS_IT-140M with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "colinglab/CLASS_IT-140M" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "colinglab/CLASS_IT-140M", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/colinglab/CLASS_IT-140M
- SGLang
How to use colinglab/CLASS_IT-140M with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "colinglab/CLASS_IT-140M" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "colinglab/CLASS_IT-140M", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "colinglab/CLASS_IT-140M" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "colinglab/CLASS_IT-140M", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use colinglab/CLASS_IT-140M with Docker Model Runner:
docker model run hf.co/colinglab/CLASS_IT-140M
Update README.md
Browse files
README.md
CHANGED
|
@@ -128,6 +128,44 @@ Use the code below to get started with the model.
|
|
| 128 |
The model has been submitted to the 2025 BabyLM Challenge – Interaction Track:
|
| 129 |
https://huggingface.co/spaces/BabyLM-community/babylm-leaderboard-2025-all-tasks
|
| 130 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 131 |
|
| 132 |
<!--
|
| 133 |
### Testing Data, Factors & Metrics
|
|
|
|
| 128 |
The model has been submitted to the 2025 BabyLM Challenge – Interaction Track:
|
| 129 |
https://huggingface.co/spaces/BabyLM-community/babylm-leaderboard-2025-all-tasks
|
| 130 |
|
| 131 |
+
## Citation
|
| 132 |
+
|
| 133 |
+
This model was introduced in the paper:
|
| 134 |
+
**“CLASS-IT: Conversational and Lecture-Aligned Small-Scale Instruction Tuning for BabyLMs”**
|
| 135 |
+
*(Capone, Bondielli & Lenci, BabyLM Challange 2025)*
|
| 136 |
+
📄 [ArXiv: 2510.25364](https://arxiv.org/abs/2510.25364)
|
| 137 |
+
|
| 138 |
+
**Cite as (BibTeX)**:
|
| 139 |
+
```
|
| 140 |
+
@inproceedings{capone-etal-2025-class,
|
| 141 |
+
title = "{CLASS}-{IT}: Conversational and Lecture-Aligned Small-Scale Instruction Tuning for {B}aby{LM}s",
|
| 142 |
+
author = "Capone, Luca and
|
| 143 |
+
Bondielli, Alessandro and
|
| 144 |
+
Lenci, Alessandro",
|
| 145 |
+
editor = "Charpentier, Lucas and
|
| 146 |
+
Choshen, Leshem and
|
| 147 |
+
Cotterell, Ryan and
|
| 148 |
+
Gul, Mustafa Omer and
|
| 149 |
+
Hu, Michael Y. and
|
| 150 |
+
Liu, Jing and
|
| 151 |
+
Jumelet, Jaap and
|
| 152 |
+
Linzen, Tal and
|
| 153 |
+
Mueller, Aaron and
|
| 154 |
+
Ross, Candace and
|
| 155 |
+
Shah, Raj Sanjay and
|
| 156 |
+
Warstadt, Alex and
|
| 157 |
+
Wilcox, Ethan Gotlieb and
|
| 158 |
+
Williams, Adina",
|
| 159 |
+
booktitle = "Proceedings of the First BabyLM Workshop",
|
| 160 |
+
month = nov,
|
| 161 |
+
year = "2025",
|
| 162 |
+
address = "Suzhou, China",
|
| 163 |
+
publisher = "Association for Computational Linguistics",
|
| 164 |
+
url = "https://aclanthology.org/2025.babylm-main.30/",
|
| 165 |
+
pages = "436--444",
|
| 166 |
+
ISBN = "TODO"
|
| 167 |
+
}
|
| 168 |
+
```
|
| 169 |
|
| 170 |
<!--
|
| 171 |
### Testing Data, Factors & Metrics
|