Error: Tensor must have a shape comprised of positive integers but got shape [1917,]

When trying to implement your react app, I keep getting this error when I run it:

index.esm.js:300 Uncaught (in promise) Error: Tensor must have a shape comprised of positive integers but got shape [1917,].
    at g (index.esm.js:300)
    at index.esm.js:592
    at Array.forEach (<anonymous>)
    at Q (index.esm.js:591)
    at kn (index.esm.js:3010)
    at Object.Dn [as tensor2d] (index.esm.js:3047)
    at index.esm.js:35762
    at index.esm.js:1622
    at t.scopedRun (index.esm.js:1628)
    at t.tidy (index.esm.js:1617)
    at Object.Ke [as tidy] (index.esm.js:2677)
    at _callee$ (index.esm.js:35761)
    at tryCatch (runtime.js:45)
    at Generator.invoke [as _invoke] (runtime.js:264)
    at Generator.prototype.<computed> [as next] (runtime.js:98)
    at asyncGeneratorStep (asyncToGenerator.js:3)
    at _next (asyncToGenerator.js:25)

I guess it has nothing to do with the react app itself, but with the google colab training phase instead (I dunno since I dont know that much of python or machine learning tho)

I've tried a couple of models (all of them with +150 images), 1 or more tags, different pics resolutions, and keep getting this error. I have also recloned the react app just to be sure I haven't touched anything

Any kind of information is appreciated.

cloud-annotations / google-colab-training

Error: Tensor must have a shape comprised of positive integers but got shape [1917,] #4