Instructions to use ehzoah/exo-imdb-sft-model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ehzoah/exo-imdb-sft-model with Transformers:

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

pipe = pipeline("text-generation", model="ehzoah/exo-imdb-sft-model")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ehzoah/exo-imdb-sft-model")
model = AutoModelForCausalLM.from_pretrained("ehzoah/exo-imdb-sft-model")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use ehzoah/exo-imdb-sft-model with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/ehzoah/exo-imdb-sft-model

SGLang

How to use ehzoah/exo-imdb-sft-model 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 "ehzoah/exo-imdb-sft-model" \
    --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": "ehzoah/exo-imdb-sft-model",
		"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 "ehzoah/exo-imdb-sft-model" \
        --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": "ehzoah/exo-imdb-sft-model",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ehzoah/exo-imdb-sft-model with Docker Model Runner:
```
docker model run hf.co/ehzoah/exo-imdb-sft-model
```

Towards Efficient Exact Optimization of Language Model Alignment

model: exo-imdb-sft-model
- Finetuned from model: gpt2-large
dataset: imdb (original stanford version)

SFT model used in the imdb experiment of the ICML'24 paper Towards Efficient Exact Optimization of Language Model Alignment.

For details of the dataset, training and inference of this model, please refer to https://github.com/haozheji/exact-optimization/blob/main/exp/imdb_exp/README.md

Downloads last month: 12

Collection including ehzoah/exo-imdb-sft-model

Efficient Exact Optimization

Collection

SFT & Reward Models used in the experiments of the ICML 2024 paper "Towards Efficient Exact Optimization of Language Model Alignment" • 2 items • Updated Jun 24, 2024

Paper for ehzoah/exo-imdb-sft-model

Towards Efficient and Exact Optimization of Language Model Alignment

Paper • 2402.00856 • Published Feb 1, 2024 • 2