Image-Text-to-Text
Transformers
ONNX
Safetensors
English
medical
chest-xray
radiology
clip
blip
multimodal
cpu
Instructions to use GAD-Research-Lab/MedicalAI-Light-Weight with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use GAD-Research-Lab/MedicalAI-Light-Weight with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="GAD-Research-Lab/MedicalAI-Light-Weight")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("GAD-Research-Lab/MedicalAI-Light-Weight", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use GAD-Research-Lab/MedicalAI-Light-Weight with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "GAD-Research-Lab/MedicalAI-Light-Weight" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "GAD-Research-Lab/MedicalAI-Light-Weight", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/GAD-Research-Lab/MedicalAI-Light-Weight
- SGLang
How to use GAD-Research-Lab/MedicalAI-Light-Weight 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 "GAD-Research-Lab/MedicalAI-Light-Weight" \ --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": "GAD-Research-Lab/MedicalAI-Light-Weight", "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 "GAD-Research-Lab/MedicalAI-Light-Weight" \ --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": "GAD-Research-Lab/MedicalAI-Light-Weight", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use GAD-Research-Lab/MedicalAI-Light-Weight with Docker Model Runner:
docker model run hf.co/GAD-Research-Lab/MedicalAI-Light-Weight
File size: 6,787 Bytes
e93bfbd | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 | import importlib
import os
import subprocess
import sys
import tempfile
from datetime import datetime
from pathlib import Path
from PIL import Image, ImageOps
DATA_DIR = Path("./data")
IMAGES_DIR = DATA_DIR / "images"
def _ensure_dep(package_name, import_name=None):
if import_name is None:
import_name = package_name
try:
return importlib.import_module(import_name)
except ImportError:
from rich.console import Console
console = Console()
console.print(f"[yellow]'{package_name}' is required for this feature.[/yellow]")
import questionary
install = questionary.confirm(f"Install {package_name} now?", default=True).ask()
if not install:
return None
console.print(f"[cyan]Installing {package_name}...[/cyan]")
subprocess.check_call([sys.executable, "-m", "pip", "install", package_name])
return importlib.import_module(import_name)
def _ensure_dirs():
IMAGES_DIR.mkdir(parents=True, exist_ok=True)
def _save_image(pil_image, prefix="capture"):
_ensure_dirs()
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S_%f")
filename = f"{prefix}_{timestamp}.jpg"
path = str(IMAGES_DIR / filename)
if pil_image.mode != "RGB":
pil_image = pil_image.convert("RGB")
pil_image.save(path, quality=95)
return path
# ββ Camera ββββββββββββββββββββββββββββββββββββββββββββββββββββββ
def capture_camera():
cv2 = _ensure_dep("opencv-python", "cv2")
if cv2 is None:
return None, "Camera capture requires opencv-python"
cap = cv2.VideoCapture(0)
if not cap.isOpened():
return None, "No camera detected (could not open index 0)"
from rich.console import Console
console = Console()
console.print("[cyan]Camera opened. Press SPACE to capture, ESC to cancel.[/cyan]")
import questionary
input("Press Enter when ready for camera preview...")
ret, frame = cap.read()
cap.release()
if not ret:
return None, "Failed to capture frame from camera"
preview_path = tempfile.mktemp(suffix="_preview.jpg")
cv2.imwrite(preview_path, frame)
preview = Image.open(preview_path)
os.unlink(preview_path)
console.print("[cyan]Image captured from camera.[/cyan]")
path = _save_image(preview, "camera")
return path, f"Captured from camera -> {path}"
# ββ File browser βββββββββββββββββββββββββββββββββββββββββββββββ
def capture_file():
try:
import tkinter as tk
from tkinter import filedialog
root = tk.Tk()
root.withdraw()
root.attributes("-topmost", True)
path = filedialog.askopenfilename(
title="Select an X-ray image",
filetypes=[
("Image files", "*.jpg *.jpeg *.png *.bmp *.tif *.tiff *.dcm"),
("All files", "*.*"),
],
)
root.destroy()
except Exception as e:
return None, f"File dialog failed: {e}"
if not path:
return None, "No file selected"
return _open_and_save(path)
# ββ Manual path entry ββββββββββββββββββββββββββββββββββββββββββ
def capture_path():
from rich.console import Console
console = Console()
console.print("[cyan]Enter the path to an X-ray image file.[/cyan]")
import questionary
path = questionary.path("Image path:").ask()
if not path:
return None, "No path entered"
return _open_and_save(path)
# ββ DICOM ββββββββββββββββββββββββββββββββββββββββββββββββββββββ
def capture_dicom(path=None):
pydicom = _ensure_dep("pydicom")
if pydicom is None:
return None, "DICOM loading requires pydicom"
np = _ensure_dep("numpy")
if np is None:
return None, "DICOM loading requires numpy"
if not path:
try:
import tkinter as tk
from tkinter import filedialog
root = tk.Tk()
root.withdraw()
root.attributes("-topmost", True)
path = filedialog.askopenfilename(
title="Select a DICOM file",
filetypes=[("DICOM files", "*.dcm"), ("All files", "*.*")],
)
root.destroy()
except Exception as e:
return None, f"File dialog failed: {e}"
if not path:
return None, "No DICOM file selected"
try:
ds = pydicom.dcmread(path)
arr = ds.pixel_array
arr = arr - arr.min()
arr = (arr / arr.max() * 255).astype(np.uint8)
if len(arr.shape) == 2:
img = Image.fromarray(arr, mode="L")
img = ImageOps.equalize(img)
else:
img = Image.fromarray(arr)
result_path = _save_image(img, "dicom")
return result_path, f"DICOM loaded from {path} -> saved as {result_path}"
except Exception as e:
return None, f"Failed to read DICOM: {e}"
# ββ Generic open + save ββββββββββββββββββββββββββββββββββββββββ
def _open_and_save(source_path):
source_path = str(source_path)
if source_path.lower().endswith(".dcm"):
return capture_dicom(source_path)
try:
img = Image.open(source_path)
path = _save_image(img, "import")
return path, f"Imported from {source_path} -> {path}"
except Exception as e:
return None, f"Failed to open image: {e}"
# ββ Top-level picker βββββββββββββββββββββββββββββββββββββββββββ
def pick_image():
from rich.console import Console
import questionary
console = Console()
method = questionary.select(
"How do you want to provide the X-ray image?",
choices=[
"Browse files on computer",
"Enter file path manually",
"Capture from camera",
"Load DICOM file",
],
pointer=">",
).ask()
result = None
if method == "Browse files on computer":
result = capture_file()
elif method == "Enter file path manually":
result = capture_path()
elif method == "Capture from camera":
result = capture_camera()
elif method == "Load DICOM file":
result = capture_dicom()
return result
if __name__ == "__main__":
path, msg = pick_image()
print(msg)
|