Invalid argument error.

[jasonyli]

I got this error while running the line (on a GPU):

embeddings = ce.get_embeddings(X_train, y_train, categorical_embedding_info=embedding_info, is_classification=True, epochs=100,batch_size=256)

Can't tell from this error what the problem is. Is it embedding #71 (indexed from 0) the 71th LabelEncoder or the 71st column in the original X_train dataframe that is causing issue?

'2 root error(s) found.\n (0) Invalid argument: indices[5,0] = 2 is not in [0, 2)\n\t [[node model/embedding_71/embedding_lookup (defined at /miniconda3/lib/python3.8/site-packages/categorical_embedder/init.py:175) ]]\n (1) Invalid argument: indices[5,0] = 2 is not in [0, 2)\n\t [[node model/embedding_71/embedding_lookup (defined at /miniconda3/lib/python3.8/site-packages/categorical_embedder/init.py:175) ]]\n\t [[ArithmeticOptimizer/ReorderCastLikeAndValuePreserving_int32_ExpandDims_23/_537]]\n0 successful operations.\n0 derived errors ignored. [Op:__inference_train_function_11843]\n\nErrors may have originated from an input operation.\nInput Source operations connected to node model/embedding_71/embedding_lookup:\n model/embedding_71/embedding_lookup/7853 (defined at /miniconda3/lib/python3.8/contextlib.py:113)\n\nInput Source operations connected to node model/embedding_71/embedding_lookup:\n model/embedding_71/embedding_lookup/7853 (defined at /miniconda3/lib/python3.8/contextlib.py:113)\n\nFunction call stack:\ntrain_function -> train_function\n'