Check failed: outputs_[index].tensor == nullptr (0x2956ae200 vs. nullptr)

[boyshell]

2021-05-09 15:33:52.796078: I tensorflow/compiler/tf2mlcompute/kernels/mlc_subgraph_op.cc:326] Compute: Failed in processing TensorFlow graph sequential/simple_rnn/while/body/1/sequential/simple_rnn/while/simple_rnn_cell/MLCSubgraphOp_2_1 with frame_id = 11685270897896072049 and iter_id = 0 with error: Internal: ExecuteMLCInferenceGraph: Failed to execute MLC inference graph. (error will be reported 5 times unless TF_MLC_LOGGING=1).
2021-05-09 15:33:52.796386: F tensorflow/core/framework/op_kernel.cc:983] Check failed: outputs[index].tensor == nullptr (0x169e91b70 vs. nullptr)
src.zip

[devnev39]

This error happens due to the numpy version. Consider changing it to 1.19.5 from 1.20 . The error may persist in model.fit() due to the simple rnn but rest of the error associated with model.predict() (in case of other models not containing rnn lstm or grus) will be resolved.

I am still waiting for this to solve but the numpy downgradation may help with other errors too.