Baptiste Roziere

Hi, Thank you. We have not really done an ablation study on the dataset size. However, the numbers you are quoting are for non deduplicated functions. We get about the...

Hi and sorry for the late answer, We created the tests by generating test cases for each problem, then running the generated tests on the ground truth. We considered the...

Hi, It may be because all 35 examples in the python file you kept are sent to the training set. Maybe train running it on the 3 python files in...

Hi. As your issue is with fastBPE, you may want to ask them if they support windows on [their repo](https://github.com/glample/fastBPE). We did not test our code on windows and you...

Hi, So it's failing when trying to consolidate the test summaries created by evosuite. You could rerun the test creation command with --rerun=True (or delete the `tests` directory in your...

Hi. Given sufficient data, I believe that you could use the same objectives as TransCoder to translate between SQL dialects.

Hi, TransCoder is not really made for pre-training a model for things like clone detection. What you are trying to do would reload only the encoder of TransCoder and fine-tune...

Hi, We definitely managed to test models with the same encoder parameters as TransCoder on CodeXGlue before. I did not test it recently, and I guess there will still be...

It's done on GPUs but we didn't really optimize the translate.py script. For instance it can only take one example at a time. If you want to translate several functions,...

Hi, I guess you are referring to the @@ characters ? These characters were created by fastBPE when breaking the tokens into subtokens. To undo the BPE, you can just...