Mobilenet does not accept batched inputs

Bug Report

Which model does this pertain to?

Mobilenet

Describe the bug

The documentation of mobilenet says that the expected input shape is (N x 3 x H x W), where N is the batch size, and H and W are expected to be at least 224.

But when downloading the mobilenetv2-7.onnx model and running a simple inference session on it, it fails when running on any batch size > 1

Reproduction instructions

System Information

OS Platform and Distribution: Ubuntu 22.04.5 LTS
ONNX version: 1.17.0
Backend/Runtime version: onnxruntime 1.21.0

Provide a code snippet to reproduce your errors.

import onnxruntime
import requests
import numpy as np

MODEL_PATH = 'mobilenet.onnx'
model_url = "https://github.com/onnx/models/raw/refs/heads/main/validated/vision/classification/mobilenet/model/mobilenetv2-7.onnx"

# Download the model
response = requests.get(model_url)
with open(MODEL_PATH, 'wb') as f:
    f.write(response.content)

# Load model
sess = onnxruntime.InferenceSession(MODEL_PATH)
input_name = sess.get_inputs()[0].name
input_shape = sess.get_inputs()[0].shape
input_type = sess.get_inputs()[0].type

print("Input shape:", input_shape)
print("Input type:", input_type)

# Create a batched input (batch size = 4)
batched_input = np.random.rand(4, 3, 224, 224).astype(np.float32)

# Try inference - should raise an error
try:
    output = sess.run(None, {input_name: batched_input})
except Exception as e:
    print("\n❌ Inference failed with batched input:")
    print(e)

Result

❌ Inference failed with batched input:
[ONNXRuntimeError] : 2 : INVALID_ARGUMENT : Got invalid dimensions for input: data for the following indices
 index: 0 Got: 4 Expected: 1
 Please fix either the inputs/outputs or the model.

Start in the linked validated/vision/classification/mobilenet/ directory and inspect the mobilenetv2-7.onnx input metadata. Run the provided Python reproduction with onnxruntime using a batch larger than one. Done means the model accepts the documented (N, 3, H, W) input and the documentation matches the verified behavior.