This solution might look very long and wierd, but trust me it doesnt throw any errors/ exceptions. It works perfect.
You could also try doing by activating the virtual env,
For example in my case, I have different environment where I use to work around all ML, NLP code here.
Activate your virtual env first,
$ source activate myML-env
which will open new environment, like below
(myML-env)$
Now within this env, hit enter after this below command to install your tflearn package only within this environment.
(myML-env)$ pip install tflearn
Don't panic if tflearn works well even in python 3.X.
(NOTE: Make sure you have tensorflow already installed in this Virtual environment, if not use the same command with the slight change [pip install tensorflow])
Now follow the below command to deactivate the current virtual env,
$ source deactivate myML-env
Now open your Anaconda >> Environments >> Select myML-env, now open your spyder application from this environment.