Instructions to use mlfoundations/tabula-8b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlfoundations/tabula-8b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="mlfoundations/tabula-8b")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mlfoundations/tabula-8b")
model = AutoModelForCausalLM.from_pretrained("mlfoundations/tabula-8b")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use mlfoundations/tabula-8b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mlfoundations/tabula-8b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mlfoundations/tabula-8b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/mlfoundations/tabula-8b

SGLang

How to use mlfoundations/tabula-8b 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 "mlfoundations/tabula-8b" \
    --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": "mlfoundations/tabula-8b",
		"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 "mlfoundations/tabula-8b" \
        --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": "mlfoundations/tabula-8b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use mlfoundations/tabula-8b with Docker Model Runner:
```
docker model run hf.co/mlfoundations/tabula-8b
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

This repository contains the TabuLa-8B (Tabular Llama-8B) model. TabuLa-8B is a foundation model for prediction (classification and binned regression) on tabular data.

TabuLa-8B is described in the paper "Large Scale Transfer Learning for Tabular Data via Language Modeling."

For more details on the model, see the paper, which includes a Model Card detailing the model architecture, training, and evaluation. TabuLa-8B was trained with rtfm, using the T4 dataset.

TabuLa-8B is built with Meta Llama 3.

Usage and Examples

You can load the model with transformers via

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mlfoundations/tabula-8b")
model = AutoModelForCausalLM.from_pretrained("mlfoundations/tabula-8b")

For more information on how to prepare data and run inference (including a demo notebook for performing inference on your data), see the examples in rtfm.

License and Terms of Use

TabuLa-8B is fine-tuned from the Llama-3 8B model. As a result, we release it under the Llama 3 license, and by using the model you agree to abide by the Llama 3 Community License Agreement and the Llama 3 Acceptable Use Policy.