Zhipu just released another OCR model: GLM-OCR, PDF 1.86 pages per second, 1 yuan can run 2000 A4 scanned images, 0.9B, can be deployed at the edge
GLM-OCR is an efficient OCR tool that processes PDF documents at a speed of 1.86 pages per second. It can process approximately 2,000 A4 scanned images for 1 yuan, and the cost is only one-tenth of traditional OCR. It supports handwriting, complex tables and multi-language mixing, is suitable for high concurrency and edge deployment, and scored a high score of 94.6 on OmniDocBench V1.5. Can be deployed through vLLM, SGLang and Ollama to reduce inference latency and computing power overhead
GLM-OCR is an efficient OCR tool that processes PDF documents at a speed of 1.86 pages per second. It can process approximately 2,000 A4 scanned images for 1 yuan, and the cost is only one-tenth of traditional OCR. It supports handwriting, complex tables and multi-language mixing, is suitable for high concurrency and edge deployment, and scored a high score of 94.6 on OmniDocBench V1.5. Can be deployed through vLLM, SGLang and Ollama to reduce inference latency and computing power overhead