Shauray Singh comments

Results 13 comments of


                                            Shauray Singh

audio classification example script RuntimeError on evaluation

Hi, @adamkatav I was not reproduce the error you mentioned in the above messages, I'm on the same versions (for the most part) as mentioned.

audio classification example script RuntimeError on evaluation

Yes, the issue you encountered with the RuntimeError could potentially be related to the Docker environment. If you are experiencing the same error in Colab as well, it suggests that...

fix type annotations for arguments in training_args

Fixing CI errors!

fix type annotations for arguments in training_args

@sgugger I don't particularly understand why this error occurs in examples_flax - ``` argparse.ArgumentError: argument --sharded_ddp: invalid typing.Union[str, bool, typing.List[transformers.trainer_utils.ShardedDDPOption], NoneType] value: '' ```

fix type annotations for arguments in training_args

@sgugger `bool` breaks `--sharded_ddp`, I think we can still maintain Boolean arguments with string itself and https://github.com/huggingface/transformers/blob/20d6b84613984f2497587a62774704882ccbeee6/src/transformers/hf_argparser.py#L168-L173 with this `--sharded_ddp` and `--fsdp` defaults to string

Training ResNet_50v1.5 on ImageNet

I'm not sure why but with GPU=0 it still trains on the CPU with just a single core. Maybe you can tell me how to train on a GPU?

Training ResNet_50v1.5 on ImageNet

Thank you @wozeparrot looks like I don't have drivers for openCL. Does using `GPU=1` or `GPU=2` make a difference?

Training ResNet_50v1.5 on ImageNet

I'm using Ubuntu right now on an Intel i5

Training ResNet_50v1.5 on ImageNet

I do have a dedicated Nvidia GPU

Training ResNet_50v1.5 on ImageNet

Training on a relatively smaller dataset with a batch size of 8 (due to memory constraints), should be reproducible by `CUDA=1 python3 ./examples/mlperf/model_train`.