I have a few questions regarding FastAI v1 I would appreciate if someone can shed some lights on them in 101 language for Dummies
1- first and the most important one, it seems that in v1 we cannot use quite a few of the pre-trained models that we could use in previous version for example models.resnexct101_64 and models.Inception_v2 are not defined is this true?
2- How we can see the the details of API’s for V1 via VSCode like what Jeremy is showing in lesson 8 for older version of fastai? what is the trick to ge the new version’s APIs?
3- How we can set “batch size” in V1? it seems that there is no such a thing when we define learner or ImageDataBunch
This is what I am receiving when try to set the batch_size =8
data = ImageDataBunch.from_folder(f’{d_path}’, ds_tfms=(tfms, []), valid =‘test’, batch_size(8))
ERROR: positional argument follows keyword argument
in general is there any minimum number of images required per category for training? I am getting also issues when I have 4 image per categories to train on.
Thanks a ton!
shahinfar