Instructions to use FastFlowLM/medgemma-4b-it-NPU2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FastFlowLM/medgemma-4b-it-NPU2 with Transformers:

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

pipe = pipeline("image-text-to-text", model="FastFlowLM/medgemma-4b-it-NPU2")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("FastFlowLM/medgemma-4b-it-NPU2", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use FastFlowLM/medgemma-4b-it-NPU2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FastFlowLM/medgemma-4b-it-NPU2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FastFlowLM/medgemma-4b-it-NPU2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/FastFlowLM/medgemma-4b-it-NPU2

SGLang

How to use FastFlowLM/medgemma-4b-it-NPU2 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 "FastFlowLM/medgemma-4b-it-NPU2" \
    --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": "FastFlowLM/medgemma-4b-it-NPU2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "FastFlowLM/medgemma-4b-it-NPU2" \
        --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": "FastFlowLM/medgemma-4b-it-NPU2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use FastFlowLM/medgemma-4b-it-NPU2 with Docker Model Runner:
```
docker model run hf.co/FastFlowLM/medgemma-4b-it-NPU2
```

ngdxzy commited on Dec 9, 2025

Commit

809e25d

1 Parent(s): 3834d00

feat: optimize vision

Browse files

Files changed (5) hide show

config.json +1 -1
dequant.xclbin +1 -1
mm.xclbin +2 -2
vision_mm.xclbin +2 -2
vision_weight.q4nx +2 -2

config.json CHANGED Viewed

@@ -32,7 +32,7 @@
     "addr_l_begin_mha": 53760,
     "addr_l_end_mha": 25088,
     "addr_kk": 45056,
-    "flm_version": "0.9.20",
     "vision_model_weight": "vision_weight.q4nx",
     "vision_mm_engine_xclbin_name": "vision_mm.xclbin",
     "vision_mha_engine_xclbin_name":"vision_attn.xclbin",

     "addr_l_begin_mha": 53760,
     "addr_l_end_mha": 25088,
     "addr_kk": 45056,
+    "flm_version": "0.9.23",
     "vision_model_weight": "vision_weight.q4nx",
     "vision_mm_engine_xclbin_name": "vision_mm.xclbin",
     "vision_mha_engine_xclbin_name":"vision_attn.xclbin",

dequant.xclbin CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:40a924d5a00bcbe983df36a2ba2bb62bb8d265cd78e310b27e5db370f0c1fcf8
 size 114059

 version https://git-lfs.github.com/spec/v1
+oid sha256:c7a392a444613b0d281f4f2d05f1f7d1d34eb26a6824ac44dbe1f871209b505f
 size 114059

mm.xclbin CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b1e88cd899e3fbef3b402ecaf2ab644f21cca2ee85dafe6e2521d89222c5a644
-size 347675

 version https://git-lfs.github.com/spec/v1
+oid sha256:aba4dbcb8b033dac2abe9a16ebf173296a7cc365c5bca6a09dee5c344ab82efe
+size 507419

vision_mm.xclbin CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72ac5cd86916bd0b77440d190db7ca7958cc94926a3bab94e86a062e1874d049
-size 387099

 version https://git-lfs.github.com/spec/v1
+oid sha256:e0b511bc64cbc43955ef78056593a1fa9848b7adcf7af43b6e6a259733de9c42
+size 516123

vision_weight.q4nx CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3d101648fb92a34137dd4b42828dc33b9daaf75e2e4ed64efd1ddd7b4742a085
-size 1844564176

 version https://git-lfs.github.com/spec/v1
+oid sha256:415579eff318b7113d568d43d5e921d9d74141eeacd0704dad0c66a8ecd911df
+size 973721768