mightyplow

Sorry for the late response. Thank you for the request. I'll look into this and I'll let you know, whether I can implement this.

Hi @rixo, thanks you for taking care of this! I created a basic example [here](https://github.com/mightyplow/svelte_hot_css_import_exmaple). You just have to install the dependencies and run `npm start` (python 3 required). It...

This would be really cool. I already tried to debug the whole process a bit but I did not know where to actually look at. But sounds like this would...

I wrapped the saving of the model in tf.engine().startScope() and tf.engine().endScope() to prevent the leaking tensor.

You have to dispose the xs and ys after the fit step. Otherwise tfjs creates new tensors in every loop step. They fill up the memory with each step if...

it should be like this ``` let xs = tf.tensor2d( a ); let ys = tf.tensor2d( b ); await model.fit(xs, ys, { epochs: 50000000, shuffle: false, verbose: 0, callbacks:{ onTrainBegin:...

Oh sorry, my fault. I didn't recognize the amount of epochs. Then you're right and it looks like some internal problem. I'll try it out and see what happens.