No support for pdf input

During the experiment of using pdf as the input, it showed
```bash
error: cannot identify image file '/home/yiyuanli/Documents/Main/OCR/DeepSeek-OCR/images/MMLUPRO.pdf'
Traceback (most recent call last):
File "/myenv/DeepSeek-OCR-master/DeepSeek-OCR-hf/run_dpsk_ocr.py", line 34, in
res = model.infer(tokenizer, prompt=prompt, image_file=image_file, output_path = output_path, base_size = 1024, image_size = 640, crop_mode = True, save_results = True, test_compress = True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/myenvi/.cache/huggingface/modules/transformers_modules/deepseek-ai/DeepSeek-OCR/59512895521e669adb87064b37e5749fe9b9f5d2/modeling_deepseekocr.py", line 745, in infer
images = load_pil_images(conversation)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/myenv/.cache/huggingface/modules/transformers_modules/deepseek-ai/DeepSeek-OCR/59512895521e669adb87064b37e5749fe9b9f5d2/modeling_deepseekocr.py", line 300, in load_pil_images
pil_img = pil_img.convert("RGB")
^^^^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'convert'
```