Instructions to use toolevalxm/ScienceGPT-TestRepo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use toolevalxm/ScienceGPT-TestRepo with Transformers:

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

pipe = pipeline("text-generation", model="toolevalxm/ScienceGPT-TestRepo")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("toolevalxm/ScienceGPT-TestRepo")
model = AutoModelForCausalLM.from_pretrained("toolevalxm/ScienceGPT-TestRepo")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use toolevalxm/ScienceGPT-TestRepo with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/toolevalxm/ScienceGPT-TestRepo

SGLang

How to use toolevalxm/ScienceGPT-TestRepo 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 "toolevalxm/ScienceGPT-TestRepo" \
    --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": "toolevalxm/ScienceGPT-TestRepo",
		"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 "toolevalxm/ScienceGPT-TestRepo" \
        --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": "toolevalxm/ScienceGPT-TestRepo",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use toolevalxm/ScienceGPT-TestRepo with Docker Model Runner:
```
docker model run hf.co/toolevalxm/ScienceGPT-TestRepo
```

ScienceGPT

1. Introduction

ScienceGPT is a specialized language model fine-tuned for scientific reasoning and knowledge. The model has been trained on extensive scientific literature and datasets, demonstrating exceptional capabilities in physics, chemistry, biology, mathematics, and earth sciences. It excels at solving complex scientific problems and explaining scientific concepts.

The model shows remarkable improvements in scientific domain understanding compared to general-purpose models. For instance, in standardized science examinations, ScienceGPT achieves 85% accuracy compared to 65% for baseline models.

2. Evaluation Results

Comprehensive Benchmark Results

	Benchmark	BaseModel	Model-v1	Model-v2	ScienceGPT
Core Sciences	Physics	0.620	0.645	0.660	0.593
	Chemistry	0.580	0.595	0.610	0.628
	Biology	0.640	0.670	0.685	0.735
	Mathematics	0.710	0.735	0.750	0.727
	Earth Science	0.590	0.615	0.630	0.661

Overall Performance Summary

ScienceGPT demonstrates strong performance across all scientific benchmark categories, with particularly notable results in mathematics and biology.

3. Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("username/ScienceGPT-TestRepo")
tokenizer = AutoTokenizer.from_pretrained("username/ScienceGPT-TestRepo")

4. License

This model is licensed under the Apache 2.0 License.

5. Contact

For questions, please contact science@example.com.

Downloads last month: 1