Instructions to use readerbench/RoSummary-large with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use readerbench/RoSummary-large with Transformers:

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

pipe = pipeline("text-generation", model="readerbench/RoSummary-large")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("readerbench/RoSummary-large")
model = AutoModelForCausalLM.from_pretrained("readerbench/RoSummary-large")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use readerbench/RoSummary-large with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/readerbench/RoSummary-large

SGLang

How to use readerbench/RoSummary-large 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 "readerbench/RoSummary-large" \
    --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": "readerbench/RoSummary-large",
		"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 "readerbench/RoSummary-large" \
        --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": "readerbench/RoSummary-large",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use readerbench/RoSummary-large with Docker Model Runner:
```
docker model run hf.co/readerbench/RoSummary-large
```

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Model card for RoSummary-large

language: - ro

RoSummary

This is a version of the RoGPT2 model trained on the AlephNews dataset for the summarization task. There are 3 trained versions, they are available on the HuggingFace Hub:

Evaluation on AlephNews

Model	Decode Method		BERTScore			ROUGE
		Precision	Recall	F1-Score	ROUGE-1	ROUGE-2	ROUGE-L
	Greedy	0.7335	0.7399	0.7358	0.3360	0.1862	0.3333
Base	Beam Search	0.7354	0.7468	0.7404	0.3480	0.1991	0.3416
	Top-p Sampling	0.7296	0.7299	0.7292	0.3058	0.1452	0.2951
	Greedy	0.7378	0.7401	0.7380	0.3422	0.1922	0.3394
Medium	Beam Search	0.7390	0.7493	0.7434	0.3546	0.2061	0.3467
	Top-p Sampling	0.7315	0.7285	0.7294	0.3042	0.1400	0.2921
	Greedy	0.7376	0.7424	0.7391	0.3414	0.1895	0.3355
Large	Beam Search	0.7394	0.7470	0.7424	0.3492	0.1995	0.3384
	Top-p Sampling	0.7311	0.7301	0.7299	0.3051	0.1418	0.2931

Acknowledgments

Research supported with Cloud TPUs from Google's TensorFlow Research Cloud (TFRC)

Downloads last month: 11