Hello, I am Paul Bustios, a master's student at University of São Paulo. I am interested in this idea for GSoC 2017 and I have a question about the Show and Tell model: will training the model be part of the GSoC project or a pre-trained model will be provided?
I also found an issue in the current implementation of the RESTful service for object recognition TIKA-1993 due to Numpy. I reported this in TIKA-2308 and I have a possible solution here https://github.com/apache/tika/commit/d29e3972c27cce421878c84224f94f259a67e84f. Please, let me know if this fix is ok. Best regards, Paul
