Short sequences work fine.Longer sequences cause GPU crashes.No code changes were made.Only input size increased.
Decode Trail Latest Questions
My RNN works fine on short sequences.When I give it longer inputs, predictions become random.Loss increases with sequence length.It feels like the model forgets earlier information.
The agent performs well in simulation.When deployed in the real world, it makes strange decisions.The physics is slightly different.Small changes lead to big failures.
The base model worked well before.After fine-tuning on new data, accuracy drops everywhere.Even old categories are misclassified.The model seems to have forgotten what it knew.
I trained a CNN to classify multiple object categories from images.The training completes without errors and the accuracy looks decent.But when I run inference, every image gets the same label.Even very different images are predicted as the ...
I am training a convolutional neural network on a custom image dataset using PyTorch.For the first few batches the loss looks normal, but suddenly it becomes NaN and never recovers.There are no crashes or stack traces, only the ...
My diagnostic CNN shows high accuracy on data from one hospital.When tested on scans from a different hospital, performance drops drastically.The disease patterns are the same.Only the scanners and imaging pipelines differ.
I trained a Keras model that gives good validation accuracy.After saving and loading it, the predictions become completely wrong.Even training samples are misclassified.Nothing crashes, but the outputs no longer make sense.
My image classifier performs very well on bright daylight photos.When images are darker or taken indoors, accuracy drops sharply.The objects are still the same.Only the lighting seems different.
My language model produces fluent responses.Even when it does not know the answer, it sounds confident.Users sometimes trust incorrect replies.There is no indication of uncertainty.