Malinda

@fchollet Below is the output of lint check. ``` Run sh shell/lint.sh [4](https://github.com/keras-team/keras/runs/7731885042?check_suite_focus=true#step:6:5) no issues with isort [5](https://github.com/keras-team/keras/runs/7731885042?check_suite_focus=true#step:6:6) no issues with flake8 [6](https://github.com/keras-team/keras/runs/7731885042?check_suite_focus=true#step:6:7) would reformat keras/utils/text_dataset_test.py [7](https://github.com/keras-team/keras/runs/7731885042?check_suite_focus=true#step:6:8) [8](https://github.com/keras-team/keras/runs/7731885042?check_suite_focus=true#step:6:9) Oh no!...

@gbaned this is done. please help me to rerun the workflow.

@fchollet Running in `black` introduces the following changes; I'm not sure if these changes should be pushed. I removed the white spaces that were causing the lint failure. Hopefully, it...

@gbaned can you please help me to rerun the workflow?

I ran `sh shell/format.sh` and pushed the changes. Hopefully this will fix the issue.

@fchollet rebased

Please check the NumPy API documentation here (https://numpy.org/doc/stable/reference/generated/numpy.linalg.multi_dot.html) If you do not trust this doc, I can run a small performance check for you. Thanks for the comment.

@piskvorky In the PR, there were two calls to `np.dot`, which we can replace with one call to `np.linalg.multi_dot`, Invoking several `np.dot` invocations is less efficient, according to the API...

I proposed this patch as an external contributor to make the code more concise and efficient. But how frequently do you use the script `svd error.py` and what effect does...

@merrymercy @ezyang I am new to this project. Could you please help me to find a reviewer to this PR?