bigbug

keras: var model = Training(); model.save_weights("your_save_path_file.h5"); is this what you want ?

I get. I can't solve this problem. the author doesn't seem to provide relevant examples.

I saw an [example](https://github.com/SciSharp/SciSharp-Stack-Examples/blob/991cfecd9e2497b3ae730330d98ec10b55812e8b/src/TensorFlowNET.Examples/ImageProcessing/CnnInYourOwnData.cs#L487) that uses another approach, by defining `Saver` Obj to save it, but this method doesn't work with keras API. `var saver = tf.train.Saver(tf.global_variables());` `saver.save(sess, ModelPath);`

After training, the memory will be automatically released. check your code

model.save_weights("your_model_path_file.h5")

just only `No space left on device`, you could free space to build.

If you just want to compile your own firmware, please fork this project and modify the configurations in the `configs` directory and trigger the corresponding `action`. Do not modify the...

Further testing set the gaming PC resolution to 1080P, AndroidTv defaults to 1080P, and the streaming display remains blurry, which doesn't seem to be caused by scaling issues.

🚀After testing, the problem finally appeared in the encoder, the processor of the device used is [Amlogic T950D4](https://www.amlogic.com/#Products/417/index.html), the chip supports 10-bit AV1/H.265/VP9/H.264 decoding. In the software set the default...