Hongkun Yu

Hi, Mingxing and I wrote the original efficientnet: https://github.com/tensorflow/tpu/blob/b24729de804fdb751b06467d3dce0637fa652060/models/official/efficientnet/efficientnet_builder.py#L31 and the model garden version should reproduce the efficientnet training. The efficientnet is trained with ``` MEAN_RGB = [0.485 * 255,...

@Zahlii I did further investigation as the following: ``` image = tf.ones((1, 224, 224, 3), dtype=tf.float32) / 2.0 inp = tf.keras.layers.Input(shape=(224, 224, 3)) layer = tfh.KerasLayer("https://tfhub.dev/tensorflow/efficientnet/b0/feature-vector/1", trainable=False)(inp) model_hub = tf.keras.models.Model(inputs=inp,...

Conflicts need to resolve. Reassign to @ByzanTine

These mentions happen in the tests. I think most of them don't require deterministic order. We can apply the optimization you mentioned. If anyone expect certain inputs and outputs, the...

Is this an official model? Could you let us know the model repo you are referring to?

@jereliu adding the author for concrete suggestions

Do you have imagenet dataset processed ready? The tutorials don't contain processed imagenet dataset tfrecords because of the license issue.

Using official/vision/beta/train.py is right. You need to provide --experiment. https://github.com/tensorflow/models/blob/master/official/vision/beta/configs/image_classification.py#L115 @yeqingli @jaeyounkim I feel we need to have basic documentation of the main models ASAP. The projects already have documentation.

I am not sure if /data/imagenet/train* is a gcp bucket path? Should it be gs://data/imagenet/train*? @arashwan

ValueError: imagenet-2012-tfrecord/train* means you are not using tfds. We place a placeholder file path here because we cannot host imagenet datasets according to its policy. Here is asking you to...