DeepSeek-OCR/configuration.json

{"framework": "pytorch", "task": "image-text-to-text", "allow_remote": true}