Only the FIRST image will be processed in batch inputs 

I use code below:

```python
    pdf_images = load_file(test_path, config={"page_range": "0-2"})
    manager = InferenceManager(method="hf")
    inputs = [BatchInputItem(image=img, prompt_type="ocr_layout") for img in pdf_images]
    results = manager.generate(inputs)
```
Length of the results is `1` and only the first image is processed. The PDF file contains 10 pages.
