Running SAM on Large Orthomosaic and Implementing Encoder-Decoder Workflow

[Ankit-Vohra]

I am currently facing challenges while trying to load and run the Segment Anything model on a large orthomosaic, which is approximately 8-10 GB in size. I would like to request your assistance in understanding the necessary steps to achieve this successfully.

Request 1: Load and Run Segment Anything Model on Large Orthomosaic I kindly request detailed guidance on how to load and execute the Segment Anything model on a large orthomosaic dataset. Specifically, I am interested in knowing the appropriate steps and configurations required to handle such a large image efficiently. If possible, it would be immensely helpful if a demo or example could be provided on how to run the Segment Anything model on a sample large orthomosaic. This would significantly aid users like me in understanding the workflow better and applying it to our own datasets.

Request 2: Implementing Encoder-Decoder Workflow Additionally, I would like to explore the implementation of an encoder-decoder workflow using the provided SAM model. I am particularly interested in learning how to pass a large image through the encoder, store the encoded vector, and then perform real-time inference using the decoder model based on the encoded vector.

Guidance Request: To achieve the encoder-decoder workflow, I request detailed steps or guidelines on how to:

Input a large image into the encoder of the SAM model. Store the encoded vector generated by the encoder. Implement the decoder model to run on top of the encoded vector in real-time for semantic segmentation tasks.