The following hyperparameters were used during training:
RESIZE_TO = 384
CROP_TO = 224
BATCH_SIZE = 64
STEPS_PER_EPOCH = 10
AUTO = tf.data.AUTOTUNE # optimise the pipeline performance
NUM_CLASSES = 5 # number of classes
SCHEDULE_LENGTH = (
500 # we will train on lower resolution images and will still attain good results
)
SCHEDULE_BOUNDARIES = [
200,
300,
400,
]
The hyperparamteres like
SCHEDULE_LENGTH
and
SCHEDULE_BOUNDARIES
are determined based on empirical results. The method has been explained in the
original paper
and in their
Google AI Blog Post
.
bit huggingface.co is an AI model on huggingface.co that provides bit's model effect (), which can be used instantly with this keras-io bit model. huggingface.co supports a free trial of the bit model, and also provides paid use of the bit. Support call bit model through api, including Node.js, Python, http.
bit huggingface.co is an online trial and call api platform, which integrates bit's modeling effects, including api services, and provides a free online trial of bit, you can try bit online for free by clicking the link below.
bit is an open source model from GitHub that offers a free installation service, and any user can find bit on GitHub to install. At the same time, huggingface.co provides the effect of bit install, users can directly use bit installed effect in huggingface.co for debugging and trial. It also supports api for free installation.