Conversation
|
Please dismiss the previous Approve, it's a mistake click. |
There was a problem hiding this comment.
I have a bunch of rephrasing comments, maybe once the other reviewers approve this PR, we can merge this and I can create a new PR with the suggested rephrasing and you folks can review it. I was trying to point them out here, but there are quite a few of them, so I stopped at two.
| @@ -0,0 +1,81 @@ | |||
| # Models continuous integration | |||
There was a problem hiding this comment.
"Models continuous integration" => "Continuous integration for models"
| This framework aims to add continuous integration to formal models based on PaddlePaddle. | ||
|
|
||
| ## Background | ||
| Currently, there are two official repositories that provide open source models based on PaddlePaddle: |
There was a problem hiding this comment.
As a supplement:
We should make standards for feeding arguments to model trainning.
e.g. :save_dir_path of 'generate_chinese_poetry' is different with model_save_dir of 'text_classification'.
Furthermore, building models bank as a python module may be helpful for implementing test framework conveniently. After pip install paddle-models-versionxxx we can use models as below:
import paddle.models
model = models.text_classification()
arguments = {'batch_size': 128, 'train_passes':500 ...}
model.setArgs(arguments)
model.train()
fixes: #521