Andrea Cossu
Andrea Cossu
Server log attached [errors.txt](https://github.com/ContinualAI/continual-learning-baselines/files/9556170/errors.txt)
nice, fingers crossed then!
Thanks @Harry-Up ! I think most of the features required to build dynamic network structures (and the strategies you mentioned) are already in Avalanche. We just need the bandwidth and...
I think in this case metrics like `StreamAccuracy` could not be implemented as they are now since they could not know when to emit the values (since `after_eval` would not...
We need authorization to access the colab. Please, check that the input size of your model matches the Omniglot input size (e.g., if you are using the `SimpleMLP` you may...
From a quick look at the code I think the mini-batches should be shuffled since `shuffle=True` by default in the dataloader creation. Can you please provide a quick script to...
Hi @dominik1102 , I agree that the results should be the same if you build the benchmark with `nc_benchmark` or `dataset_benchmark`. Can you post here a small script reproducing the...
Code looks fine, I am wondering if the difference is only in the metric names when you use task labels. Can you print here the output of the TextLogger when...
After a quick look I think the problem is related to the targets of the dataset of each experience. You can see it by printing them with this code: ```python...
1. In class-incremental scenarios you do not have task label available, hence you either work with a single-headed model or you work with a multi-head model but you have to...