Instructions to use inverse-scaling/opt-30b_eval with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use inverse-scaling/opt-30b_eval with Transformers:

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

pipe = pipeline("text-generation", model="inverse-scaling/opt-30b_eval")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("inverse-scaling/opt-30b_eval")
model = AutoModelForCausalLM.from_pretrained("inverse-scaling/opt-30b_eval", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use inverse-scaling/opt-30b_eval with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "inverse-scaling/opt-30b_eval"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "inverse-scaling/opt-30b_eval",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/inverse-scaling/opt-30b_eval

SGLang

How to use inverse-scaling/opt-30b_eval 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 "inverse-scaling/opt-30b_eval" \
    --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": "inverse-scaling/opt-30b_eval",
		"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 "inverse-scaling/opt-30b_eval" \
        --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": "inverse-scaling/opt-30b_eval",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use inverse-scaling/opt-30b_eval with Docker Model Runner:
```
docker model run hf.co/inverse-scaling/opt-30b_eval
```

opt-30b_eval

Commit History

Jan 2, 2023

13:01
Add verifyToken field to verify evaluation results are produced by Hugging Face's automatic model evaluator (#17)⁠

MicPie +1: autoevaluator

f1d5f8c

Nov 14, 2022

20:46
Add evaluation results on the mathemakitten--winobias_antistereotype_test_v5 config and test split of mathemakitten/winobias_antistereotype_test_v5 (#16)⁠

MicPie +1: autoevaluator

3e7fdf5
20:45
Add evaluation results on the mathemakitten--winobias_antistereotype_test_v5 config and test split of mathemakitten/winobias_antistereotype_test_v5 (#15)⁠

MicPie +1: autoevaluator

a8d9390
20:45
Add evaluation results on the mathemakitten--winobias_antistereotype_test_v5 config and test split of mathemakitten/winobias_antistereotype_test_v5 (#14)⁠

MicPie +1: autoevaluator

9ef7973

Oct 9, 2022

18:05
Update README.md

MicPie

5f79120
18:01
Update README.md

MicPie

f87aad5
18:00
Add evaluation results on the inverse-scaling--quote-repetition config and train split of inverse-scaling/quote-repetition (#4)⁠

MicPie +1: autoevaluator

1931c38

Oct 8, 2022

13:38
Add evaluation results on the inverse-scaling--NeQA config and train split of inverse-scaling/NeQA (#3)⁠

MicPie +1: autoevaluator

3b28414
12:47
Update README.md

MicPie

3ba68b7

Oct 7, 2022

20:18
Add evaluation results on the inverse-scaling--41 config and train split of inverse-scaling/41 (#2)⁠

MicPie +1: autoevaluator

8339ba1

Oct 6, 2022

09:36
cp opt-30b

Michael Pieler

857e680

Oct 5, 2022

15:06
initial commit

MicPie

5501e01

Commit History

Jan 2, 2023

Add verifyToken field to verify evaluation results are produced by Hugging Face's automatic model evaluator (#17)⁠

Nov 14, 2022

Add evaluation results on the mathemakitten--winobias_antistereotype_test_v5 config and test split of mathemakitten/winobias_antistereotype_test_v5 (#16)⁠

Add evaluation results on the mathemakitten--winobias_antistereotype_test_v5 config and test split of mathemakitten/winobias_antistereotype_test_v5 (#15)⁠

Add evaluation results on the mathemakitten--winobias_antistereotype_test_v5 config and test split of mathemakitten/winobias_antistereotype_test_v5 (#14)⁠

Oct 9, 2022

Update README.md

Update README.md

Add evaluation results on the inverse-scaling--quote-repetition config and train split of inverse-scaling/quote-repetition (#4)⁠

Oct 8, 2022

Add evaluation results on the inverse-scaling--NeQA config and train split of inverse-scaling/NeQA (#3)⁠

Update README.md

Oct 7, 2022

Add evaluation results on the inverse-scaling--41 config and train split of inverse-scaling/41 (#2)⁠

Oct 6, 2022

cp opt-30b

Oct 5, 2022

initial commit