Instructions to use OpenVINO/phi-2-int4-ov with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenVINO/phi-2-int4-ov with Transformers:

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

pipe = pipeline("text-generation", model="OpenVINO/phi-2-int4-ov", device_map="auto")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OpenVINO/phi-2-int4-ov")
model = AutoModelForCausalLM.from_pretrained("OpenVINO/phi-2-int4-ov", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OpenVINO/phi-2-int4-ov with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenVINO/phi-2-int4-ov"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenVINO/phi-2-int4-ov",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/OpenVINO/phi-2-int4-ov

SGLang

How to use OpenVINO/phi-2-int4-ov 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 "OpenVINO/phi-2-int4-ov" \
    --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": "OpenVINO/phi-2-int4-ov",
		"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 "OpenVINO/phi-2-int4-ov" \
        --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": "OpenVINO/phi-2-int4-ov",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use OpenVINO/phi-2-int4-ov with Docker Model Runner:
```
docker model run hf.co/OpenVINO/phi-2-int4-ov
```

sylwia-kuros commited on Jul 12, 2024

Commit

b7a6445

verified ·

1 Parent(s): f17adae

Update README.md

Browse files

Files changed (1) hide show

README.md +3 -3

README.md CHANGED Viewed

@@ -2,9 +2,9 @@
 license: mit
 license_link: https://choosealicense.com/licenses/mit/
 ---
-# -int4-ov
 * Model creator: [Microsoft](https://huggingface.co/microsoft)
- * Original model: [phi-2](https://huggingface.co/microsoft/phi-2)
 ## Description
 This is [phi-2](https://huggingface.co/microsoft/phi-2) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/2024/documentation/openvino-ir-format.html) (Intermediate Representation) format with weights compressed to INT4 by [NNCF](https://github.com/openvinotoolkit/nncf).
@@ -14,7 +14,7 @@ This is [phi-2](https://huggingface.co/microsoft/phi-2) model converted to the [
 Weight compression was performed using `nncf.compress_weights` with the following parameters:
 * mode: **int4_sym**
-* ratio: **1**
 * group_size: **128**
 For more information on quantization, check the [OpenVINO model optimization guide](https://docs.openvino.ai/2024/openvino-workflow/model-optimization-guide/weight-compression.html).

 license: mit
 license_link: https://choosealicense.com/licenses/mit/
 ---
+# phi-2-int4-ov
 * Model creator: [Microsoft](https://huggingface.co/microsoft)
+* Original model: [phi-2](https://huggingface.co/microsoft/phi-2)
 ## Description
 This is [phi-2](https://huggingface.co/microsoft/phi-2) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/2024/documentation/openvino-ir-format.html) (Intermediate Representation) format with weights compressed to INT4 by [NNCF](https://github.com/openvinotoolkit/nncf).
 Weight compression was performed using `nncf.compress_weights` with the following parameters:
 * mode: **int4_sym**
+* ratio: **1.0**
 * group_size: **128**
 For more information on quantization, check the [OpenVINO model optimization guide](https://docs.openvino.ai/2024/openvino-workflow/model-optimization-guide/weight-compression.html).