Benjamin Schiller

Got the same problem...so the only real solution (except for manually setting the desired capacity after each CA pod restart) is to leave desired at least at 1...?

@aceeric did you find a solution for this? I'm having similar trouble when using an AWS EFS Access point as workspace: ``` 2022/06/14 14:40:40 warning: unsuccessful cred copy: ".gitconfig" from...

Oh, I completely forgot to mention that. Yes, I use a batch size of 64. This behavior only applies if batching is used.

The samples are just ordered by character length and then batched, so they still may vary within a batch (but much less than before). The speed up just comes from...

Hey @pommedeterresautee, sorry for the long wait - I was on a holiday trip. I based my script on your demo scripts but I cannot disclose the model and/or dataset....

I noticed in the post that the speedup between TensorRT (FP16) and TensorRT (INT-8) is not as much as you @pommedeterresautee were observing, so I also finally tested the QAT...

Is there some update on this matter? We have the same problem, unfortunately.

@alex2awesome I described adding own data [here](https://github.com/UKPLab/mdl-stance-robustness#add-you-own-datasets) and how to run the training [here](https://github.com/UKPLab/mdl-stance-robustness#run-experiments-and-evaluate). Can you elaborate what is missing to use the repo? @Kaushal-13 The fine-tuned model weights are...

Hi @rudra0713, 1. Yes, you are right, there. The dataset specific dense layers for the classification are not included. 2. No, this won't be a problem. You will have to...

Hi @rudra0713, the model won't check the labels, it will only leverage the index you set. It should be fine setting different indices for the same label if you use...