Michael Kowalski

The accuracy improves from 99.29% to 99.41% by using the batch mean and batch var during training. Those are just single runs but the lower one is pretty far outside...

Looking at the code on the master now, have you set `validation_size = 0` in input_data.py so that all examples get used for training? I have fixed the error pointed...

The final accuracy was 99.33%. I got that result on 2 training runs. If I put my typo back (`batch_normalization(z_pre_l, m_l, mean, var)`) the model does train to 99.41. My...

I wonder if the remaining difference is the implementation of Adam in Tensorflow vs Blocks. [Blocks has different defaults and also an extra decay term](http://blocks.readthedocs.org/en/latest/api/algorithms.html#blocks.algorithms.Adam) that is not available in...

Apparently using a placeholder for a conditional in a tensorflow graph does not work with a simple `if` - see http://stackoverflow.com/a/35833133/728291. Using the placeholder with `tf.cond`, as done [in batch...

This can be done with `_Raw.randomShuffle` if you only shuffle on the first dimension and don’t need the indices anywhere else. Otherwise it can be done with `_Raw.gatherV2` where you...

An example with `_Raw.randomShuffle`: import TensorFlow let x = Tensor(shape: [5, 2], scalars: Array(0 ..< 10)) let y = _Raw.randomShuffle(value: x, seed: 1, seed: 2) print(y) // [[8, 9], [2,...

I should have added that importing 'x10_training_loop' works as expected using the MacOS v0.9 toolchain or the June 12 toolchain for MacOS. So I think it is either specific to...

Hi @8bitmp3. Yes, my reduced example with `Device.trainingDevices` can be worked around or I can train in x10 using other code. I demonstrated the problem with `.trainingDevices` only because it...

I have been counting parameters based on how variables used to be updated in optimisers: extension Layer { var parameterCount: Int { let floatKeyPaths = self.recursivelyAllWritableKeyPaths(to: Tensor.self) let doubleKeyPaths =...