AttributeError: 'list' object has no attribute 'items'

[parhameftekhar]

I'm trying to train the MBT model on my own dataset. I get the following error. Any help is appreciated.

Traceback (most recent call last):
File "/home/eftekhar/anaconda3/lib/python3.9/runpy.py", line 197, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/home/eftekhar/anaconda3/lib/python3.9/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/home/eftekhar/models/scenic/scenic/projects/mbt/main.py", line 49, in
app.run(main=main)
File "/home/eftekhar/anaconda3/lib/python3.9/site-packages/scenic/app.py", line 65, in run
app.run(functools.partial(_run_main, main=main))
File "/home/eftekhar/anaconda3/lib/python3.9/site-packages/absl/app.py", line 308, in run
_run_main(main, args)
File "/home/eftekhar/anaconda3/lib/python3.9/site-packages/absl/app.py", line 254, in _run_main
sys.exit(main(argv))
File "/home/eftekhar/anaconda3/lib/python3.9/site-packages/scenic/app.py", line 100, in _run_main
main(rng=rng, config=FLAGS.config, workdir=FLAGS.workdir, writer=writer)
File "/home/eftekhar/models/scenic/scenic/projects/mbt/main.py", line 39, in main
trainer.train(
File "/home/eftekhar/anaconda3/lib/python3.9/site-packages/scenic/projects/mbt/trainer.py", line 425, in train
gflops) = mbt_train_utils.initialize_model(
File "/home/eftekhar/anaconda3/lib/python3.9/site-packages/scenic/projects/mbt/train_utils.py", line 83, in initialize_model
for modality_name, spec in input_spec.items():
AttributeError: 'list' object has no attribute 'items'

The config file also is:

r"""Multimodal sound classification on the balanced (mini) AudioSet.

"""

import ml_collections

AUDIOSET_TRAIN_SIZE = 20361

def get_config():
"""Returns the base experiment configuration."""
config = ml_collections.ConfigDict()
config.experiment_name = 'mbt_balanced_audioset_classification'

config.dataset_configs = ml_collections.ConfigDict()
config.dataset_configs.base_dir = '/home/eth/tfrecords_builder/tmp/generated_dataset'
config.dataset_configs.tables = {
'train': 'train',
'validation': 'valid',
'test': 'test',
}
config.dataset_configs.examples_per_subset = {
'train': 4189,
'validation': 898,
'test': 898
}
config.dataset_configs.num_classes = 20
config.data_dtype_str = 'float32'
config.dataset_name = 'video_tfrecord_dataset'
config.dataset_configs.modalities = ('spectrogram', 'rgb')
config.dataset_configs.return_as_dict = False
config.dataset_configs.num_frames = 32
config.dataset_configs.stride = 2
config.dataset_configs.num_spec_frames = 8
config.dataset_configs.spec_stride = 1

config.dataset_configs.spec_mean = 1.102
config.dataset_configs.spec_stddev = 2.762

config.dataset_configs.min_resize = 256
config.dataset_configs.crop_size = 224
config.dataset_configs.spec_shape = (100, 128)

config.dataset_configs.one_hot_labels = True
config.dataset_configs.zero_centering = True

config.dataset_configs.do_multicrop_test = True
config.dataset_configs.log_test_epochs = 4
config.dataset_configs.num_test_clips = 4
config.dataset_configs.test_batch_size = 8 # Needs to be num_local_devices
config.multicrop_clips_per_device = 2

config.dataset_configs.augmentation_params = ml_collections.ConfigDict()
config.dataset_configs.augmentation_params.do_jitter_scale = True
config.dataset_configs.augmentation_params.scale_min_factor = 0.9
config.dataset_configs.augmentation_params.scale_max_factor = 1.33
config.dataset_configs.augmentation_params.prob_scale_jitter = 1.0
config.dataset_configs.augmentation_params.do_color_augment = True
config.dataset_configs.augmentation_params.prob_color_augment = 0.8
config.dataset_configs.augmentation_params.prob_color_drop = 0.1

config.dataset_configs.prefetch_to_device = 2

config.dataset_configs.spec_augment = True
config.dataset_configs.spec_augment_params = ml_collections.ConfigDict()
config.dataset_configs.spec_augment_params.freq_mask_max_bins = 48
config.dataset_configs.spec_augment_params.freq_mask_count = 1
config.dataset_configs.spec_augment_params.time_mask_max_frames = 48
config.dataset_configs.spec_augment_params.time_mask_count = 4
config.dataset_configs.spec_augment_params.time_warp_max_frames = 1.0
config.dataset_configs.spec_augment_params.time_warp_max_ratio = 0
config.dataset_configs.spec_augment_params.time_mask_max_ratio = 0

config.model_name = 'mbt_multilabel_classification'
config.model = ml_collections.ConfigDict()
config.model.modality_fusion = ('spectrogram', 'rgb')
config.model.use_bottleneck = True
config.model.test_with_bottlenecks = True
config.model.share_encoder = False
config.model.n_bottlenecks = 4
config.model.fusion_layer = 8
config.model.hidden_size = 768
config.model.patches = ml_collections.ConfigDict()
config.model.attention_config = ml_collections.ConfigDict()
config.model.attention_config.type = 'spacetime'
config.model.num_heads = 12
config.model.mlp_dim = 3072
config.model.num_layers = 12
config.model.representation_size = None
config.model.classifier = 'gap'
config.model.attention_dropout_rate = 0.
config.model.dropout_rate = 0.
config.model_dtype_str = 'float32'

config.model.temporal_encoding_config = ml_collections.ConfigDict()
config.model.temporal_encoding_config.method = '3d_conv'
config.model.patches.size = [16, 16, 2]
config.model.temporal_encoding_config.kernel_init_method = 'central_frame_initializer'
config.model.temporal_encoding_config.n_sampled_frames = 4 # Unused here.

config.trainer_name = 'mbt_trainer'
config.optimizer = 'momentum'
config.optimizer_configs = ml_collections.ConfigDict()
config.l2_decay_factor = 0
config.max_grad_norm = 1
config.label_smoothing = 0.3
config.num_training_epochs = 50
config.batch_size = 64
config.rng_seed = 0
config.mixup = ml_collections.ConfigDict()
config.mixup.alpha = 0.5
config.mixmod = False
config.model.stochastic_droplayer_rate = 0.3

config.init_from = ml_collections.ConfigDict()
config.init_from.model_config = None
disable=line-too-long
config.init_from.checkpoint_path = '/home/eth/models/scenic/scenic/projects/mbt/vit_base'
config.init_from.checkpoint_format = 'scenic'
config.init_from.model_config = ml_collections.ConfigDict()
config.init_from.model_config.model = ml_collections.ConfigDict()
config.init_from.model_config.model.classifier = 'token' # Specify if this is 'token' or 'gap'. pylint: disable=line-too-long
config.init_from.restore_positional_embedding = True
config.init_from.restore_input_embedding = True
config.init_from.positional_embed_size_change = 'resize_tile'

steps_per_epoch = AUDIOSET_TRAIN_SIZE // config.batch_size
total_steps = config.num_training_epochs * steps_per_epoch
config.lr_configs = ml_collections.ConfigDict()
config.lr_configs.learning_rate_schedule = 'compound'
config.lr_configs.factors = 'constant * cosine_decay * linear_warmup'
config.lr_configs.warmup_steps = 2.5 * steps_per_epoch
config.lr_configs.steps_per_cycle = total_steps
config.lr_configs.base_learning_rate = 5e-1

config.write_summary = True
config.checkpoint = True # Do checkpointing.
config.debug_train = False # Debug mode during training.
config.debug_eval = False # Debug mode during eval.
config.checkpoint_steps = 500 # Checkpoint more frequently than a val epoch.
return config

[huangfei00]

I encountered the same problem as you did. Did you manage to solve it?

[huangfei00]

emmm..... Also, for the preprocessing of the Audioset dataset, could you provide some reference code?

[parhameftekhar]

@huangfei00 No I gave up