tensorflow - TFLearn - Evaluate a model -

i using tflearn alexnet sample own dataset.

next want perform classification on test data , determine accuracy of model.

1. tflearn api provides methods model.predict() , model.evaluate().
2. model.predict() gives prediction result each image in test data set. how can use result accuracy?
3. model.evaluate() gives accuracy score on test data. there way accuracy each batch well?

below responses:

2. you can calculate accuracy comparing predicted classes against effective ones when using model.predict()
3. no, aware of. not sure useful use case either: interested on overall accuracy dataset/partition evaluating.