Instructions to use tiny-random/phi-4-multimodal with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiny-random/phi-4-multimodal with Transformers:

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

pipe = pipeline("text-generation", model="tiny-random/phi-4-multimodal", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("tiny-random/phi-4-multimodal", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use tiny-random/phi-4-multimodal with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiny-random/phi-4-multimodal"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiny-random/phi-4-multimodal",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tiny-random/phi-4-multimodal

SGLang

How to use tiny-random/phi-4-multimodal 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 "tiny-random/phi-4-multimodal" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiny-random/phi-4-multimodal",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "tiny-random/phi-4-multimodal" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiny-random/phi-4-multimodal",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tiny-random/phi-4-multimodal with Docker Model Runner:
```
docker model run hf.co/tiny-random/phi-4-multimodal
```

yujiepan commited on Feb 28, 2025

Commit

ca89336

verified ·

1 Parent(s): 0bb056e

Upload folder using huggingface_hub

Browse files

Files changed (2) hide show

README.md +1 -3
model.safetensors +2 -2

README.md CHANGED Viewed

@@ -182,9 +182,7 @@ with open(Path(save_folder, 'phi4mm/vision_siglip_navit.py'), 'w') as f:
 sys.path.append(str(Path(save_folder)))
 from phi4mm.modeling_phi4mm import Phi4MMForCausalLM
 print(Phi4MMForCausalLM)  # ensure imported
-torch.set_default_dtype(torch.bfloat16)
-model = Phi4MMForCausalLM(config)
-torch.set_default_dtype(torch.float32)
 set_seed(42)
 with torch.no_grad():

 sys.path.append(str(Path(save_folder)))
 from phi4mm.modeling_phi4mm import Phi4MMForCausalLM
 print(Phi4MMForCausalLM)  # ensure imported
+model = Phi4MMForCausalLM(config).to(torch.bfloat16)
 set_seed(42)
 with torch.no_grad():

model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9f810845d6961348ac0b6ea813434359c36b89c1546f378497b5cbce0fb79bdd
-size 6557984

 version https://git-lfs.github.com/spec/v1
+oid sha256:a47f5caa9957c85df39e16d6c7e781fdd3db271bf7f36872eeaa4c51864dca3e
+size 6551432