I created a model on Tensor Flow. I tried uploading it to the PersonalImageClassifier1 Extension, but it said the file was to large. Is there anything I can to do reduce the size of the file or upload the existing file?
I expect you are using the advice here>Personal Image Classifier: Part 1 to create a training file. If so, you are downloading images using a web cam. The only way you can control the fize size of the images in this instance is to set your web cam to provide the lowest resolution possible is my guess. My web cam cannot do that 
but yours might be adjustable. 
Of course, you can reduce the number of training images. Or load small images using the old model tool https://classifier.appinventor.mit.edu./oldpic/ That might work.
Could you please tell me how many images you added in how many labels?