Google Dataflow unsupported SDK version 0.5.5 error 400

While doing the Google Cloud Transfer learning flowers examples using my own data images, I have issues with launching the Dataflow with the correct version and getting an unsupported SDK error. It seems like the program somehow still download the old verison of Dataflow SDK instead of new ones. However, I have tried following upgrades before executing code :  

*sudo pip install --upgrade -r requirements.txt

sudo pip install pip==9.0.3 ( to fix other error) 

Code I am trying to use is Google cloud pre-processing using dataflow :

https://github.com/GoogleCloudPlatform/cloudml-samples (source)

code :

Error log:

Collecting google-cloud-dataflow==0.5.5 Using cached https://files.pythonhosted.org/packages/18/18/e7d7a32f917d47bbcc7e3d5c9b60138a7d51adf04d155e0d400108915e79/google-cloud-dataflow-0.5.5.tar.gz Saved /tmp/tmp17oW97/google-cloud-dataflow-0.5.5.tar.gzSuccessfully downloaded google-cloud-dataflow

File "/home/zeyihe/.local/lib/python2.7/site-packages/apitools/base/py/base_api.py", line 604, in __ProcessHttpResponse http_response, method_config=method_config, request=request)apitools.base.py.exceptions.HttpBadRequestError: HttpError accessing <https://dataflow.googleapis.com/v1b3/projects/nexgenai-224306/jobs?alt=json>: response: <{'status': '400', 'content-length': '342', 'x-xss-protection': '1; mode=block', 'x-content-type-options': 'nosniff', 'transfer-encoding': 'chunked', 'vary': 'Origin, X-Origin, Referer', 'server': 'ESF', '-content-encoding': 'gzip', 'cache-control': 'private', 'date': 'Sun, 06 Jan 2019 13:11:24 GMT', 'x-frame-options': 'SAMEORIGIN', 'content-type': 'application/json; charset=UTF-8'}>, content <{ "error": { "code": 400, "message": "(717cdaa221ed221e): The workflow was automatically rejected by the service because it uses an unsupported SDK version 0.5.5. For a list of supported SDK versions, see: \nhttps://cloud.google.com/dataflow/support#support-status-for-dataflow-sdk-releases", "status": "INVALID_ARGUMENT" }}