[Train/docs] Deduplicate Train examples & move from ray.air to ray.train
#29667
Conversation
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Yard1
requested review from
richardliaw,
krfricke,
xwjiang2010,
amogkam,
matthewdeng,
maxpumperla and
a team
as code owners
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
| # args = ["--smoke-test"] | ||
| # ) | ||
|
|
||
| # py_test( |
There was a problem hiding this comment.
Is this one already being tested in test_examples?
There was a problem hiding this comment.
test_gpu_examples, will add a comment
|
|
||
| # py_test( | ||
| # name = "tensorflow_autoencoder_example", # REGRESSION | ||
| # size = "medium", |
There was a problem hiding this comment.
this autoencoder example is still under the ray/air directory.
Let's keep this commented out pytest in the air BUILD file.
There was a problem hiding this comment.
Actually, the .py file got moved here. The jupyter notebook is in ray/air, and that is not what this entry is referring to.
| ) | ||
|
|
||
| py_test( | ||
| name = "horovod_cifar_pbt_example", |
There was a problem hiding this comment.
do we need an __init__.py in the ray/train/examples/horovod?
There was a problem hiding this comment.
I think we do. We may run into issues with imports (and we are importing functions from the examples) without it. It's present there.
| @@ -1,7 +1,7 @@ | |||
| import argparse | |||
|
|
|||
| from ray.air.config import RunConfig, ScalingConfig | |||
| from ray.train.examples.torch_fashion_mnist_example import train_func | |||
| from ray.train.examples.pytorch.torch_fashion_mnist_example import train_func | |||
| from ray.train.torch import TorchTrainer | |||
| from ray.air.callbacks.mlflow import MLflowLoggerCallback | |||
There was a problem hiding this comment.
should this example be moved into pytorch?
There was a problem hiding this comment.
I guess it is using pytorch, but on the other hand, it's supposed to show off MLFlow first and foremost and pytorch is only an implementation detail here. I'd just leave it where it is for better visibility.
Signed-off-by: Antoni Baum <[email protected]>
|
Failing tests are also failing on master. Going to merge. |
….train` (ray-project#29667) This PR moves Train examples from ray.air.examples to ray.train.examples, making sure that duplicate examples are removed and the folder structure is preserved. Signed-off-by: Antoni Baum <[email protected]> Signed-off-by: Weichen Xu <[email protected]>
Signed-off-by: Antoni Baum [email protected]
Why are these changes needed?
This PR moves Train examples from
ray.air.examplestoray.train.examples, making sure that duplicate examples are removed and the folder structure is preserved.Related issue number
Closes #26536
Checks
git commit -s) in this PR.scripts/format.shto lint the changes in this PR.