Value Error Occured in training dataset with Jasper Model

[secrisa11]

python ./bin/main.py model=jasper train=jasper_train train.dataset_path=$DATASET_PATH train.transcripts_path=$TRANSCRIPTS_PATH ./bin/main.py:175: UserWarning: 'audio/fbank' is validated against ConfigStore schema with the same name. This behavior is deprecated in Hydra 1.1 and will be removed in Hydra 1.2. See https://hydra.cc/docs/next/upgrades/1.0_to_1.1/automatic_schema_matching for migration instructions. main() [2021-07-09 15:36:37,587][kospeech.utils][INFO] - audio: audio_extension: pcm sample_rate: 16000 frame_length: 20 frame_shift: 10 normalize: true del_silence: true feature_extract_by: kaldi time_mask_num: 4 freq_mask_num: 2 spec_augment: true input_reverse: false transform_method: fbank n_mels: 80 freq_mask_para: 18 model: architecture: jasper teacher_forcing_ratio: 1.0 teacher_forcing_step: 0.01 min_teacher_forcing_ratio: 0.9 dropout: 0.3 bidirectional: false joint_ctc_attention: false max_len: 400 version: 10x5 train: dataset: kspon dataset_path: /home/suresoft/KoSpeech-1.3/dataset/kspon transcripts_path: /home/suresoft/KoSpeech-1.3/dataset/kspon/transcripts.txt output_unit: character batch_size: 32 save_result_every: 1000 checkpoint_every: 5000 print_every: 10 mode: train num_workers: 4 use_cuda: true init_lr_scale: 0.01 final_lr_scale: 0.05 max_grad_norm: 400 weight_decay: 0.001 seed: 777 resume: false optimizer: novograd reduction: sum init_lr: 0.001 final_lr: 0.0001 peak_lr: 0.001 warmup_steps: 0 num_epochs: 10 lr_scheduler: tri_stage_lr_scheduler

[2021-07-09 15:36:37,755][kospeech.utils][INFO] - Operating System : Linux 4.9.201-tegra [2021-07-09 15:36:37,756][kospeech.utils][INFO] - Processor : aarch64 [2021-07-09 15:36:37,797][kospeech.utils][INFO] - device : NVIDIA Tegra X1 [2021-07-09 15:36:37,797][kospeech.utils][INFO] - CUDA is available : True [2021-07-09 15:36:37,798][kospeech.utils][INFO] - CUDA version : 10.2 [2021-07-09 15:36:37,798][kospeech.utils][INFO] - PyTorch version : 1.6.0 [2021-07-09 15:36:37,827][kospeech.utils][INFO] - split dataset start !! [2021-07-09 15:36:41,561][kospeech.utils][INFO] - Applying Spec Augmentation... [2021-07-09 15:36:45,168][kospeech.utils][INFO] - Applying Spec Augmentation... Error executing job with overrides: ['model=jasper', 'train=jasper_train', 'train.dataset_path=/home/suresoft/KoSpeech-1.3/dataset/kspon', 'train.transcripts_path=/home/suresoft/KoSpeech-1.3/dataset/kspon/transcripts.txt'] Traceback (most recent call last): File "./bin/main.py", line 175, in main() File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/main.py", line 53, in decorated_main config_name=config_name, File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/_internal/utils.py", line 368, in _run_hydra lambda: hydra.run( File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/_internal/utils.py", line 214, in run_and_report raise ex File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/_internal/utils.py", line 211, in run_and_report return func() File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/_internal/utils.py", line 371, in overrides=args.overrides, File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/_internal/hydra.py", line 110, in run _ = ret.return_value File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/core/utils.py", line 233, in return_value raise self._return_value File "/home/suresoft/miniforge3/envs/KoSpeech_Py36/lib/python3.6/site-packages/hydra/core/utils.py", line 160, in run_job ret.return_value = task_function(task_cfg) File "./bin/main.py", line 170, in main last_model_checkpoint = train(config) File "./bin/main.py", line 99, in train epoch_time_step, trainset_list, validset = split_dataset(config, config.train.transcripts_path, vocab) File "/home/suresoft/KoSpeech-1.3/kospeech/data/data_loader.py", line 306, in split_dataset audio_extension=config.audio.audio_extension File "/home/suresoft/KoSpeech-1.3/kospeech/data/data_loader.py", line 67, in init self.shuffle() File "/home/suresoft/KoSpeech-1.3/kospeech/data/data_loader.py", line 102, in shuffle self.audio_paths, self.transcripts, self.augment_methods = zip(*tmp) ValueError: not enough values to unpack (expected 3, got 0)

[sooftware]

Hi @Daeyeop-Kim. This repository is archived. Further development is underway here.

[pvodopija]

Had the same problem and managed to fix it like this. Go to file kospeech/data/data_loader.py and in function shuffle change the following:

def shuffle(self):
        """ Shuffle dataset """
        tmp = list( zip( self.audio_paths, self.transcripts, self.augment_methods ) )
        random.shuffle( tmp )

        # This kinda works.
        for i, x in enumerate( tmp ):
            self.audio_paths[i] = x[0]
            self.transcripts[i] = x[1]
            self.augment_methods[i] = x[2]
        
        # This doesn't work.
        # self.audio_paths, self.transcripts, self.augment_methods = zip( *tmp )