Tensorflow.GradientTape(loss, convOutputs) return none

Question

I am trying to retrieve the gradient for one of the layers of my deep learning problem but my Tensorflow.GradientTape() seems not working and only getting none as output. may I ask why my Tensorflow.GradientTape() only return none? Below is my code:

with tf.GradientTape() as tape:
        # cast the image tensor to a float-32 data type, pass the
        # image through the gradient model, and grab the loss
        # associated with the specific class index
  inputs = tf.cast(img_array, tf.float32)
  (convOutputs, predictions) = gradModel(inputs)
  loss = predictions[0]
grads = tape.gradient(loss, convOutputs)
print('grads: ', grads)

The convOutputs output:

tf.Tensor(
[[[[-6.50294571e+01 -4.58692093e+01 -1.21771477e+02 -4.86483383e+01
    -4.03000793e+01 -4.47039261e+01 -3.82719078e+01  3.48307991e+01
     4.30690117e+01 -7.36239548e+01 -4.47292366e+01 -2.25696030e+01
    -5.37582512e+01 -2.76310825e+01 -6.89400940e+01 -3.52958183e+01
    -3.95952339e+01 -6.38968506e+01 -2.50861855e+01 -5.69738884e+01
     3.72166214e+01 -5.65787430e+01 -2.73997631e+01  1.14628124e+01
    -6.55105515e+01 -4.89888687e+01 -7.03046341e+01 -5.48777657e+01
    -6.88002777e+00 -3.06381378e+01 -7.07140732e+01 -3.77145309e+01]
   [-7.54676666e+01 -6.00736237e+01 -1.28173172e+02 -6.00169525e+01
    -9.04914856e+01 -5.46973114e+01 -4.90329475e+01  3.39109764e+01
     4.66986609e+00 -8.24787750e+01 -3.25911865e+01 -5.62109795e+01
    -6.04496994e+01 -4.63624191e+01 -7.25583038e+01 -4.51388702e+01
    -5.94483299e+01 -5.90684357e+01 -7.34841232e+01 -6.95459137e+01
     2.62296143e+01 -8.91154556e+01 -7.20792465e+01 -2.05469131e+01
    -6.55375748e+01 -5.54658661e+01 -7.32153015e+01 -9.26992798e+01
    -3.99403381e+01 -4.10292320e+01 -6.38874626e+01 -4.65183105e+01]
   [-1.81520477e+02 -8.58971024e+01 -1.83252670e+02 -1.59161270e+02
    -1.69194611e+02 -1.17689140e+02 -5.95196724e+01  1.45689621e+01
    -3.20590935e+01 -7.52816086e+01 -8.10653839e+01 -1.56222626e+02
    -1.00091446e+02 -1.75923462e+02 -1.20677704e+02 -8.94266891e+01
    -9.98111115e+01 -6.04290428e+01 -1.71505966e+02 -1.11467499e+02
     1.17656879e+01 -1.44185013e+02 -1.68774277e+02 -1.10262825e+02
    -6.82244720e+01 -9.06527634e+01 -9.47591782e+01 -1.24716576e+02
    -5.97154350e+01 -7.39844131e+01 -8.39862671e+01 -6.67463379e+01]
   [-2.02750793e+02 -1.11701836e+02 -2.31550232e+02 -2.28375671e+02
    -1.90685211e+02 -1.49189346e+02 -5.43388901e+01 -1.06953392e+02
    -1.08836563e+02 -1.79691040e+02 -2.08681839e+02 -1.74694946e+02
    -1.56589874e+02 -2.09311905e+02 -1.02196709e+02 -1.07891243e+02
    -1.38184296e+02 -1.67173294e+02 -1.05253479e+02 -1.78938950e+02
     1.50398626e+01 -1.94940002e+02 -1.99048111e+02 -1.50519348e+02
    -8.67778168e+01 -1.18944138e+02 -8.32716293e+01 -1.63217850e+02
    -7.73772888e+01 -1.40492661e+02 -1.59470184e+02 -9.42923813e+01]
   [-1.60187195e+02 -1.04745750e+02 -1.89076584e+02 -2.32906769e+02
    -1.35969604e+02 -1.10200859e+02 -4.13491631e+01 -8.15036316e+01
    -7.62036819e+01 -2.23744186e+02 -1.93137466e+02 -1.54521271e+02
    -1.32532852e+02 -1.66396255e+02 -1.00753769e+02 -1.77784119e+02
    -8.59329910e+01 -2.65502563e+02 -7.01343155e+01 -1.97808884e+02
     8.65319347e+00 -1.87622223e+02 -1.69195908e+02 -1.40977982e+02
    -1.91329132e+02 -1.16586494e+02 -8.00257492e+01 -1.50202347e+02
    -1.09122505e+02 -1.24707924e+02 -1.77550446e+02 -6.93529510e+01]]

  [[-6.66161270e+01 -2.64440804e+01 -1.28140366e+02 -6.08246040e+01
    -3.14667835e+01 -2.95994892e+01 -1.43016663e+01  2.27176514e+01
     7.90086222e+00 -5.30740623e+01 -4.93246841e+01 -1.58894873e+01
    -3.50432091e+01 -2.57912960e+01 -9.59140320e+01 -3.46913376e+01
    -2.13604717e+01 -6.77545929e+01  1.14354506e+01 -6.55926590e+01
     4.73260689e+01 -4.23489799e+01 -2.39433880e+01 -2.42672491e+00
    -4.37202301e+01 -4.05773163e+01 -4.15097198e+01 -4.44734917e+01
     3.15404987e+01 -2.40143623e+01 -4.35698318e+01 -3.19632702e+01]
   [-6.61377716e+01 -2.90448055e+01 -1.14564606e+02 -3.94671402e+01
    -4.19694672e+01 -3.25295906e+01 -1.34672260e+01  2.65013885e+01
    -2.64687634e+01 -4.29406700e+01 -2.47901440e+01 -2.73481178e+01
    -4.35592003e+01 -3.96362877e+01 -7.21442871e+01 -4.19899673e+01
    -2.78450775e+01 -4.29825249e+01 -1.76962662e+01 -5.75513611e+01
     2.97128105e+01 -5.48190498e+01 -3.41783180e+01 -1.06625862e+01
    -4.20104752e+01 -5.06555405e+01 -4.39076805e+01 -5.24307747e+01
     7.58871126e+00 -5.40041542e+00 -3.43171997e+01 -2.91319466e+01]
   [-1.51534653e+02 -4.91730728e+01 -1.40873367e+02 -8.98630676e+01
    -1.00344437e+02 -6.77407684e+01 -3.38970490e+01 -1.12597027e+01
    -5.42256737e+01 -7.82623291e+01 -4.71805420e+01 -9.08375244e+01
    -9.65660248e+01 -1.27125603e+02 -8.76733093e+01 -8.06944199e+01
    -6.38962364e+01 -6.03005676e+01 -7.25920792e+01 -8.43904877e+01
     3.71763725e+01 -9.26620865e+01 -1.01217278e+02 -6.47124100e+01
    -5.39962387e+01 -7.23599014e+01 -7.42565536e+01 -8.00809250e+01
    -1.52286158e+01 -9.93980598e+00 -5.45536537e+01 -5.17784653e+01]
   [-1.42126617e+02 -8.55980911e+01 -1.73071243e+02 -1.30776672e+02
    -1.40732788e+02 -9.65427780e+01 -6.02343941e+01 -7.24563293e+01
    -7.11780701e+01 -1.59952820e+02 -1.36139114e+02 -1.45440216e+02
    -1.43853790e+02 -1.34124771e+02 -7.12768250e+01 -1.27354332e+02
    -6.51488113e+01 -1.47200394e+02 -3.54433975e+01 -1.14576134e+02
     1.29357481e+00 -1.29225601e+02 -1.19338219e+02 -4.26249466e+01
    -8.78280640e+01 -9.07064056e+01 -1.27298775e+02 -9.98622284e+01
    -2.28426075e+01 -4.23718109e+01 -1.25820114e+02 -4.81174431e+01]
   [-1.34760330e+02 -6.83024292e+01 -1.61202469e+02 -1.72546783e+02
    -6.90318375e+01 -6.66799622e+01 -4.47756042e+01 -6.17167244e+01
    -4.14773369e+01 -1.81487854e+02 -1.08990417e+02 -1.30153198e+02
    -7.56170807e+01 -1.01191025e+02 -7.46376038e+01 -1.51523376e+02
    -2.58549709e+01 -2.28963791e+02 -1.84856987e+01 -1.05832977e+02
    -9.57096863e+00 -1.18948875e+02 -4.48866730e+01 -2.72828884e+01
    -1.50047195e+02 -9.94590302e+01 -1.28323502e+02 -9.25699539e+01
    -8.30309753e+01 -3.41541939e+01 -1.50688248e+02 -4.07999115e+01]]

  [[-1.00298004e+02 -2.79989662e+01 -1.32166733e+02 -6.45445786e+01
    -3.51304588e+01 -5.10005875e+01 -7.67053223e+00  1.19501438e+01
     4.05785980e+01 -6.21512718e+01 -5.85583649e+01 -4.74961662e+00
    -1.77192383e+01 -2.45590611e+01 -8.27424011e+01 -3.69177208e+01
    -3.20581551e+01 -7.62170944e+01  1.18218031e+01 -6.60456772e+01
     7.74116821e+01 -6.08136864e+01 -3.19929838e+00 -1.79827766e+01
    -6.26097069e+01 -3.99175911e+01 -4.57050591e+01 -5.30993996e+01
     4.19713058e+01 -1.56706533e+01 -3.52655678e+01 -4.44060211e+01]
   [-6.71766357e+01 -1.99459629e+01 -1.08029373e+02 -4.70797577e+01
    -2.80885086e+01 -3.77616730e+01 -4.33005524e+00  1.50996113e+01
    -1.31076050e+01 -2.96133347e+01 -3.47656784e+01 -2.83319988e+01
    -1.29751530e+01 -3.78589592e+01 -5.78071976e+01 -2.67772484e+01
    -2.92880154e+01 -5.14851608e+01  3.70119238e+00 -5.88222771e+01
     3.57400436e+01 -5.97146835e+01 -3.34626465e+01 -3.40979652e+01
    -4.92773895e+01 -3.99031754e+01 -3.41446915e+01 -5.27044716e+01
     2.24157677e+01 -9.65095806e+00 -4.09354019e+01 -2.40360851e+01]
   [-1.18293770e+02 -1.87678947e+01 -1.08181755e+02 -6.39685440e+01
    -6.83273926e+01 -3.71158714e+01 -2.26710453e+01  1.65345211e+01
    -4.70239601e+01 -6.95715942e+01 -3.56492691e+01 -4.37086372e+01
    -4.54568138e+01 -8.97499466e+01 -7.86928101e+01 -4.69689178e+01
    -4.89040413e+01 -4.06424751e+01 -3.60207443e+01 -6.92472687e+01
     3.04923820e+01 -6.80088577e+01 -7.76435318e+01 -4.71260033e+01
    -5.98833389e+01 -6.50003891e+01 -3.36851768e+01 -7.36452484e+01
     3.27237964e+00  1.94829674e+01 -6.37408562e+01 -2.98937397e+01]
   [-1.13721970e+02 -7.47574768e+01 -1.63777847e+02 -1.20889580e+02
    -1.03018990e+02 -8.19046707e+01 -3.88016052e+01 -1.83455067e+01
    -4.98403244e+01 -1.52323517e+02 -9.14826584e+01 -8.55913010e+01
    -5.89244232e+01 -9.98255234e+01 -1.59403591e+01 -1.13324051e+02
    -4.90616417e+01 -1.03549080e+02 -2.14299011e+01 -7.84435577e+01
     5.24579506e+01 -1.06978035e+02 -9.29921188e+01 -2.38253403e+01
    -9.27440491e+01 -8.24802322e+01 -1.18578148e+02 -6.86495972e+01
     5.57220268e+00 -2.30071869e+01 -1.03968422e+02 -3.36276093e+01]
   [-1.06895256e+02 -6.46759109e+01 -1.64552689e+02 -1.42266663e+02
    -6.96146927e+01 -6.89720535e+01 -3.24379082e+01 -3.56290588e+01
    -6.13865328e+00 -2.10536072e+02 -8.95202713e+01 -1.07036301e+02
    -3.21375923e+01 -9.44600372e+01 -2.76476593e+01 -1.22160149e+02
    -4.64336967e+01 -1.89343719e+02  2.99339218e+01 -7.94682541e+01
     6.00469742e+01 -9.69546356e+01 -5.70380898e+01  6.74482644e-01
    -1.13635376e+02 -6.04484253e+01 -1.14417030e+02 -7.08537674e+01
    -4.66607056e+01 -2.90317440e+01 -1.17086098e+02 -2.80717144e+01]]

  [[-7.83961029e+01 -4.38561897e+01 -1.03443848e+02 -9.40255508e+01
    -3.09591274e+01 -4.97879868e+01 -1.44933367e+01 -2.72768936e+01
     5.47082405e+01 -6.17440834e+01 -5.46072311e+01  7.51116848e+00
    -2.14971275e+01 -2.36911411e+01 -8.05374908e+01 -1.74307842e+01
    -3.85354881e+01 -7.29558640e+01  5.22867699e+01 -4.20192680e+01
     1.06591675e+02 -7.68303757e+01 -2.56950951e+01 -6.26316261e+00
    -7.05988312e+01 -3.82594414e+01 -3.07214603e+01 -4.05800438e+01
     1.65160046e+01 -4.31133728e+01 -2.53330040e+01 -4.16086960e+01]
   [-7.24018707e+01 -3.27168808e+01 -1.01690887e+02 -6.46226730e+01
    -4.26335297e+01 -4.82178726e+01 -1.75713825e+01 -4.01015434e+01
    -2.63595343e+00 -4.57209206e+01 -4.32891846e+01 -2.57844715e+01
    -2.66282654e+01 -6.09465599e+01 -7.22618408e+01 -8.12949371e+00
    -3.26540184e+01 -6.04958878e+01  1.33619709e+01 -6.63120575e+01
     5.14195976e+01 -6.31467056e+01 -3.81864166e+01 -6.02723351e+01
    -6.76874237e+01 -5.61192169e+01 -4.03058281e+01 -5.76764412e+01
     1.28933249e+01 -3.68484421e+01 -4.19150467e+01 -3.41297264e+01]
   [-1.22748901e+02 -4.61498985e+01 -1.28551773e+02 -9.43741379e+01
    -7.48201675e+01 -5.83885574e+01 -3.20087624e+01 -1.52792969e+01
    -3.05973015e+01 -7.32163696e+01 -6.41964340e+01 -3.29736862e+01
    -5.65552483e+01 -9.43927612e+01 -8.17922516e+01 -3.76888580e+01
    -4.11721992e+01 -5.56034851e+01 -4.12347260e+01 -7.53574982e+01
     2.35608025e+01 -6.99987411e+01 -7.46448898e+01 -6.00826569e+01
    -5.61940918e+01 -5.54073029e+01 -7.97652359e+01 -6.80532532e+01
     9.61763000e+00 -3.30432439e+00 -7.00994492e+01 -4.23612213e+01]
   [-1.19826340e+02 -7.43413925e+01 -1.96550430e+02 -1.52753296e+02
    -1.12129784e+02 -8.36444931e+01 -2.50004578e+01 -1.45406742e+01
    -6.15699615e+01 -1.09175713e+02 -8.56576691e+01 -3.89918022e+01
    -8.88123932e+01 -1.11232529e+02 -1.83842106e+01 -1.08175301e+02
    -6.90916061e+01 -8.43128357e+01 -5.95601082e+01 -7.40951843e+01
     2.05149765e+01 -9.82719421e+01 -1.05462479e+02 -6.37452507e+00
    -8.69417191e+01 -9.53863297e+01 -1.39556534e+02 -8.50241165e+01
     4.38243103e+01 -2.91210365e+01 -9.05255280e+01 -3.84593887e+01]
   [-1.08696198e+02 -7.12351685e+01 -1.77165665e+02 -1.53624878e+02
    -1.10382088e+02 -8.29420090e+01 -2.00975266e+01 -2.81666107e+01
    -2.10142918e+01 -1.47295349e+02 -6.15075798e+01 -4.70545387e+01
    -7.01287613e+01 -8.32409515e+01 -3.66536026e+01 -1.26987968e+02
    -8.16168137e+01 -1.78946960e+02  4.95766211e+00 -8.77789154e+01
     4.49512444e+01 -9.48104172e+01 -8.23697662e+01  4.50058842e+00
    -1.07119164e+02 -8.19478455e+01 -1.23385155e+02 -9.19227371e+01
    -2.51898909e+00 -4.86064644e+01 -9.84637756e+01 -3.62767830e+01]]

  [[-9.16252747e+01 -5.75831375e+01 -1.55861252e+02 -1.39771072e+02
    -9.48103638e+01 -7.93823166e+01 -4.24372520e+01 -2.72077541e+01
     2.17317028e+01 -1.10957077e+02 -5.61391678e+01  1.11820184e-01
    -5.62331047e+01 -6.86520844e+01 -7.68999634e+01 -3.85947266e+01
    -6.22300453e+01 -9.73161011e+01  4.00664940e+01 -3.34659081e+01
     8.90209045e+01 -1.20408669e+02 -1.04610985e+02 -5.84791565e+01
    -4.26876373e+01 -6.18165665e+01 -3.97359047e+01 -4.60310097e+01
     3.35782585e+01 -5.28617821e+01 -7.07039261e+01 -4.85386887e+01]
   [-1.07146255e+02 -6.03385544e+01 -1.86442093e+02 -1.27499496e+02
    -8.39790268e+01 -7.01283035e+01 -3.62089233e+01 -6.78185577e+01
    -2.84582615e+01 -7.75467453e+01 -5.00115128e+01 -2.60956268e+01
    -6.16244164e+01 -1.13885254e+02 -8.00021667e+01 -7.34867249e+01
    -5.71540680e+01 -1.02708046e+02 -3.76270332e+01 -8.81984100e+01
     3.82776299e+01 -1.12780403e+02 -4.73329735e+01 -1.07194603e+02
    -5.37505684e+01 -6.14227524e+01 -5.49794273e+01 -5.63342400e+01
     2.41549397e+01 -4.48131561e+01 -9.79351883e+01 -5.80604134e+01]
   [-1.75651917e+02 -7.04204178e+01 -1.70451218e+02 -1.52089050e+02
    -8.85784225e+01 -9.09223709e+01 -3.65536575e+01 -9.53903427e+01
    -5.55574646e+01 -7.43143311e+01 -1.02717918e+02 -7.23222504e+01
    -6.67450714e+01 -1.65396225e+02 -1.11778915e+02 -9.25357513e+01
    -7.12982712e+01 -1.07670708e+02 -1.18757072e+02 -1.12558533e+02
    -2.18426204e+00 -1.19183937e+02 -7.36508179e+01 -9.09444046e+01
    -7.04857635e+01 -6.74856644e+01 -1.29483688e+02 -4.29243050e+01
     3.26767540e+00 -3.33818779e+01 -1.13437317e+02 -6.23838768e+01]
   [-1.74321289e+02 -8.85020828e+01 -1.63709534e+02 -1.72335083e+02
    -1.36558685e+02 -1.07339119e+02 -3.85947189e+01 -7.53417816e+01
    -7.46018677e+01 -1.05396896e+02 -1.09803688e+02 -5.64882278e+01
    -4.60955162e+01 -1.75642471e+02 -8.42182922e+01 -1.03485435e+02
    -1.02019661e+02 -1.33778183e+02 -1.17473511e+02 -8.04779663e+01
     1.19213972e+01 -1.42490707e+02 -1.36451492e+02 -2.75964699e+01
    -1.05291977e+02 -9.01456985e+01 -1.30564651e+02 -6.75157547e+01
     3.95124016e+01 -7.90159302e+01 -1.20788155e+02 -4.93909836e+01]
   [-1.69786209e+02 -8.77394409e+01 -1.69143021e+02 -1.56078812e+02
    -1.27199379e+02 -1.01377495e+02 -2.38635807e+01 -2.72201252e+01
     9.66639423e+00 -1.15257111e+02 -7.98624344e+01 -4.44125977e+01
    -5.50779610e+01 -1.69039276e+02 -7.61957245e+01 -1.41374710e+02
    -1.16698959e+02 -2.38188675e+02  2.11598282e+01 -6.37767525e+01
     1.40245256e+01 -1.48823059e+02 -1.71909729e+02 -3.53447037e+01
    -1.05243088e+02 -1.14340263e+02 -1.05114594e+02 -8.82040939e+01
     1.95453491e+01 -9.96810837e+01 -8.80124054e+01 -5.64831314e+01]]]], shape=(1, 5, 5, 32), dtype=float32)

The loss output

tf.Tensor([[1.0000000e+00 1.2751160e-17 5.6624083e-21]], shape=(1, 3), dtype=float32)

The grads output from Tensorflow.GradientTape is

None

Answer 1

I think I have found the reason, it probably is because I am building a multitasking model but my predictions output's track is different with the track of the layer I want to retrieve the gradient