| `loss_fn` cannot be a Keras `Loss` object, pass a… | exception | error | |
| the structure of `anchor_sizes` must be a tuple, list, or… | exception | error | |
| centered_crop is only supported when… | exception | error | |
| Initializer not supported | exception | error | |
| Only support dictionary decoder_output. | exception | error | |
| boxlist must be a BoxList | exception | error | |
| Dataset must specify a path for the data files. | exception | error | |
| `dataset_or_fn` should be either callable or an instance of… | exception | error | |
| Expected rank 4 input, was | exception | error | |
| Task not found | exception | error | |
| type: is not a valid key! | exception | error | |
| `slice_by_treatment` must be set to `False` when `y_pred`… | exception | error | |
| Unknown optimizer | exception | error | |
| Unsupported optimizer type | exception | error | |
| decoder's output should be a dict,but got | validation | error | |
| epochs_between_evals > 1 not supported for file based… | exception | error | |
| The builder does not support a global batch size with more… | exception | error | |
| Unknown type | exception | error | |
| The min_level must be >= 1, but | exception | error | |
| Currently only supports RetinaNetHead. | exception | error | |
| Only support binary relation in restriction. | exception | error | |
| When passing a mask tensor to TNTransformerExpandCondense… | exception | error | |
| is not a valid mode. | exception | error | |
| Pre-batch ( ) size is not equal to batch size ( ) | exception | error | |
| Unknown model name | exception | error | |
| add_tpu_function_alias is only allowed for input_type of… | exception | error | |
| The mode is not implemented | exception | error | |
| The _feature_encoder layer must be specified if the… | exception | error | |
| TpuBatchNormalization does not support fused=True. | exception | error | |
| Image format is invalid | exception | error | |
| `mask` not of type np.uint8 | exception | error | |
| Stem type not supported. | exception | error | |
| The key ` ` does not exist. | exception | error | |
| The output of the given metric must either be a `tf.Tensor`… | exception | error | |
| [ ] must be an integer, got . | exception | error | |
| decoder_layer should be built. | validation | error | |
| `maskrcnn.Parser` only supports `pad = True`. | exception | error | |
| `sequence_output` must be in decoder output. | validation | error | |
| values and is_treatment must be tensors of shapes (D0, D1… | exception | error | |
| crop_type must be one of | exception | error | |
| Custom training loop on GPUs is not implemented. | exception | error | |
| File output is a directory, will not save outputs to file. | exception | error | |
| Full loss computation is not yet supported. | exception | error | |
| If `use_type_embeddings` is True, then… | validation | error | |
| Invalid augmentation_name | exception | error | |
| TransformerEncoderBlock expects a three-dimensional input… | exception | error | |
| Got unsupported tensor type for feature | exception | error | |
| When negatives are in between matched and unmatched… | exception | error | |
| y_pred must be of type `TwoTowerTrainingOutputs`… | exception | error | |
| Attribute not found in model outputs. | exception | error | |
| Detection module not implemented for | exception | error | |
| length of vocab_sizes | exception | error | |
| The BUILDER returns an unexpected instance. The… | exception | error | |
| The key is internally reserved. Can not be overridden. | exception | error | |
| Unknown `attn_type` . | exception | error | |
| must be a non-empty string. | exception | error | |
| Input should have rank | exception | error | |
| The length of attribute_heads should be 1 or 3, found | exception | error | |
| The model type is not defined | exception | error | |
| translations should have rank 1 or 2. | exception | error | |
| Invalid dataset received. Received | exception | error | |
| len(stddev) must match the number of channels | exception | error | |
| When passing a mask tensor to TransformerXLBlock, the mask… | exception | error | |
| All features from the feature_names set must be tensors… | exception | error | |
| collection path at position is never registered. Please… | exception | error | |
| input_path and weights must both be a Config to use… | exception | error | |
| Unsupported string type | exception | error | |
| All slicing values in the slicing spec must be one of… | exception | error | |
| data_folder must be set as the downloaded folder path. | validation | error | |
| filter_size_scale is not greater than zero. | exception | error | |
| global_step not found in checkpoint | validation | error | |
| The last dimension of predicted scores should be divisible… | exception | error | |
| When passing a mask tensor to TransformerLayer, the mask… | exception | error | |
| must be a list of exactly two integers, got . | exception | error | |
| `global_step` must be a `tf.Variable`. | exception | error | |
| If `sequence_output` is not in encoder… | validation | error | |
| The first layer size should be multiple of 2! | exception | error | |
| The argument must be a tuple of integers. Received | exception | error | |
| `segmentation_decoder` needs to be provided for Panoptic… | validation | error | |
| output_shape must be a 1-D Tensor of 2 elements… | exception | error | |
| `model` must be provided if using… | exception | error | |
| `slice_by_treatment` must be False when y_pred is a… | exception | error | |
| Unsupported attention type | exception | error | |
| optimizer is not an object of tf_keras.optimizers.Optimizer | exception | error | |
| The key ` ` is internally reserved. Can not be overridden. | exception | error | |
| Keras model.fit on TPUs is not implemented. | exception | error | |
| encoded_boxes.shape[-1] is | exception | error | |
| Please do not set eval_dataset. Progressive training relies… | exception | error | |
| Angles should have a rank 0 or 1. | exception | error | |
| Only support expanding the first or the last dimension. Got | exception | error | |
| PanopticSegmentationModule module not implemented for | validation | error | |
| Unsupported learning rate type | exception | error | |
| cuda_visible_devices must be a string or integer, got | exception | error | |
| The key does not exist in . To extend the existing keys… | exception | error | |
| boxes.shape[-1] is , but must be 4. | exception | error | |
| Unsupported relation in restriction. | exception | error | |
| Model directory must be specified. | exception | error | |
| The key ` ` does not exist. To extend the existing keys… | exception | error | |
| Not valid params: param_set= | exception | error | |
| Unknown value for shape_for_attention. | validation | error | |