Create app.py

app.py

@@ -0,0 +1,48 @@
+import gradio as gr
+from transformers import AutoModel, AutoTokenizer
+from PIL import Image
+import torch
+import os
+
+model_name = "deepseek-ai/DeepSeek-OCR"
+
+print("Loading tokenizer...")
+tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+
+print("Loading model...")
+model = AutoModel.from_pretrained(
+    model_name,
+    trust_remote_code=True,
+    _attn_implementation="flash_attention_2",
+    use_safetensors=True
+).cuda().to(torch.bfloat16).eval()
+
+def ocr_infer(image):
+    # Save temporary image
+    image.save("input.png")
+
+    # Standard OCR prompt
+    prompt = "<image>\nFree OCR."
+
+    result = model.infer(
+        tokenizer,
+        prompt=prompt,
+        image_file="input.png",
+        output_path=".",
+        base_size=1024,
+        image_size=640,
+        crop_mode=True,
+        save_results=False
+    )
+    
+    return result["text"]
+
+demo = gr.Interface(
+    fn=ocr_infer,
+    inputs=gr.Image(type="pil"),
+    outputs="text",
+    title="DeepSeek-OCR API (HuggingFace)",
+    description="Upload an image and get OCR text using DeepSeek-OCR"
+)
+
+demo.launch()