Hello author, I am very happy to learn from your open source InsightFace project.
I recently found that the project has updated the function of recognizing age and gender.
This is also a function that I am very much looking forward to and I would like to express my thanks again.
I have a question that I saw in deploy/test.py
parser.add_argument('--model', default='', help='path to load model.')
parser.add_argument('--ga-model', default='', help='path to load model.')
In previous versions, the model parameter of test.py was the r34 model.
Why was the model parameter empty in test.py this time?
And this time the ga-model has been added.
It seems that I can't find the download of the ga-model model.
I do not know if it is convenient for you to provide a download of ga-model?
Looking forward to your reply.
Good luck.
@nttstar
I found these in the code history
parser.add_argument('--age-model', default='../models2/model-r34-age/model,0', help='path to load model.')
parser.add_argument('--gender-model', default='../models2/model-r18-gender/model,0', help='path to load model.')
I think there should be a ga-model and it is still integrating.
Hope to provide ga-model download after the integration is completed.
Thanks.
nttstar just posted a ga model!
@glorioushedgehog oh my god, thank you very much for reminding me
parser.add_argument('--age-model', default='../models2/model-r34-age/model,0', help='path to load model.')
Doo you know what does "0" mean in parser.add_argument('--age-model', default='../models2/model-r34-age/model,0', help='path to load model.')?
0 denotes the number of epochs your model need to be trained.
That is if you are using a pre-trained model, and you don't want to further train (fine-tune) it...then just pass 0 to use the pretrained model as it is.
Most helpful comment
nttstar just posted a ga model!
https://github.com/deepinsight/insightface/issues/272