ghostljj/deepseek-ocr:54661d79 | Run with an API on Replicate

You're looking at a specific version of this model. Jump to the model overview.

ghostljj /deepseek-ocr:54661d79

Input schema

The fields you can use to run this model with an API. If you don’t give a value for a field its default value will be used.

Field	Type	Default value	Description
mode	None	gundam	Processing mode: tiny (512×512), small (640×640), base (1024×1024), large (1280×1280), or gundam (adaptive 640+1024)
task_type	None	markdown	Task type: markdown (convert to markdown), ocr (standard OCR), free_ocr (OCR without layout), parse_figure (parse figures/diagrams), describe (detailed description), or locate (locate objects by reference)
image	string		Input image URL or path for OCR
custom_prompt	string		Custom prompt (optional, overrides task_type if provided)

Output schema

The shape of the response you’ll get when you run this model with an API.

Schema

{'properties': {'coordinates': {'items': {},
                                'title': 'Coordinates',
                                'type': 'array'},
                'extracted_images': {'items': {},
                                     'title': 'Extracted Images',
                                     'type': 'array'},
                'image': {'title': 'Image'},
                'markdown': {'title': 'Markdown', 'type': 'string'},
                'markdown_path': {'title': 'Markdown Path'},
                'text': {'title': 'Text', 'type': 'string'},
                'text_path': {'title': 'Text Path'}},
 'required': ['text', 'coordinates'],
 'title': 'PredictionOutput',
 'type': 'object'}