I’m new to AI and was training a Cat detector as a project. I’m currently working on Google Colab and have done initial training and have the best weight that I’m using for further training. The problem I’m running into is that my new data set is about 9500 images and the time estimate it’s giving till it’s finished is roughly 400 hours. Is there anything I can change to get the hours to drop? I only have 1 class, and this is the current cfg im using:
batch is 64
subdivisions are 16
max_batches are 150000
steps are 120000 to 135000
All other settings are default except for the filters and classes which have all been changed correctly.
Please let me know if you want any more information. Any help is greatly appreciated!
Ive tried messing with the batch size, max batch size, subdivitions and steps but I havent had mich success.
Neel Dahake is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.