Open deyang2000 opened 2 weeks ago
Hi, there. I'm looking into fine-tuning methods for SAM2. My current training conditions are 2_3090 (24G) and 8_A100 (40G). May I ask the author in what environment did you train? Do I need to buy an upgrade package on Colab to use the A100?
Hi there, sorry for the late reply. This repository is just a test that I followed based on the following article: Train and Fine-Tune Segment Anything 2 (SAM) in 60 Lines of Code. In my case, I’m using a T4 GPU and it’s running fine, so I don’t think you need to buy an upgrade package on Colab.
Hi, there. I'm looking into fine-tuning methods for SAM2. My current training conditions are 23090 (24G) and 8A100 (40G). May I ask the author in what environment did you train? Do I need to buy an upgrade package on Colab to use the A100?