seb.

@PGijsbers please have a look at this one. It's a few months old, but now that `master` has all the fixes from `stable-v2`, I started to merge small pending PRs....

Hi @franchuterivera, no i don't plan to add lock support for this, and here is why. The global "results/results.csv" is just provided for convenience, as it is aggregating the results...

of course, if you think that sth like `filelock` could be useful, feel free to propose a change, your contributions are always appreciated. I personally mainly rely on the `scores/results.csv`...

Reopening this, this seems to occur much more often that I thought so will try to add a fil lock on that aggregate `results.csv` file.

@alanwilter which framework are you trying to use exactly? Detection and usage of GPUs will depends on the framework. > I see that requirements files don't have pytorch or tensorflow....

Sorry for late reply (was off for a month): @PGijsbers is right for H2O, it will detect GPUs and use them for xgboost. I don't know how it works with...

@PGijsbers no problem, ticket was created weeks ago. Just wanted to assign it given that you're currently working on it.

could you try to add ```bash set -e ``` on top of `frameworks/shared/setup.sh` ? it should ensure that every framework's setup will exit on the first failed command

there's probably additional work needed to get this aborted correctly in containers and in AWS though... (setup done separately there)

@lujiaying, you found the reason: > user_dir is not uploaded to S3, because user_dir is in app_dir if you put your `user_dir` outside the `amlb` folder structure, it should work....