Training fails after 223 epochs

[JRMeyer]

>>> Sadam1195
[June 25, 2020, 11:11am]

Hi, I started training with

> python distribute.py slash --config_path config.json slash | tee training.log

After training for around 223 epochs, it breaks and gave this out in my
training.log

> EVALUATION slash
> warning: audio amplitude out of range, auto clipped. slash
> slash | slash > Synthesizing test sentences slash
> warning: audio amplitude out of range, auto clipped. slash
> warning: audio amplitude out of range, auto clipped. slash
> warning: audio amplitude out of range, auto clipped. slash
> -- slash > EVAL PERFORMANCE slash
> slash | slash > avg_postnet_loss: 0.17037 (+0.00013) slash
> slash | slash > avg_decoder_loss: 0.25991 (-0.00214) slash
> slash | slash > avg_stopnet_loss: 0.06618 (+0.00142) slash
> slash | slash > avg_align_error: 0.35647 (-0.00567) slash
> slash | slash > avg_ga_loss: 0.01053 (-0.00001)
>
> EPOCH: 223/1000
>
> Number of output frames: 3
>
> TRAINING (2020-06-24 11:14:35) slash
> ! Run is kept in .../LJSpeech/ljspeech-June-24-2020_01+02AM-8f8ba5e slash
> slash ['train.py', '--continue_path=', '--restore_path=',
> '--config_path=config.json', '--group_id=group_2020_06_24-010230',
> '--rank=0' slash ]

I thought maybe my VM might have rebooted so I ran again with
checkpoints

> python train.py slash --config_path config.json slash --restore_path
> /workspace/LJSpeech/ljspeech-June-24-2020_01+02AM-8f8ba5e/checkpoint_50000.pth.tar
> slash | tee training.log

But after 219 epochs again my training stopped and gave this error

> EPOCH: 219/1000
>
> Number of output frames: 2
>
> TRAINING (2020-06-25 06:52:31) slash
> ! Run is kept in .../LJSpeech/ljspeech-June-24-2020_12+50PM-8f8ba5e slash
> Traceback (most recent call last): slash
> File 'train.py', line 676, in slash
> main(args) slash
> File 'train.py', line 591, in main slash
> global_step, epoch) slash
> File 'train.py', line 148, in train slash
> for num_iter, data in enumerate(data_loader): slash
> File
> '/opt/conda/lib/python3.7/site-packages/torch/utils/data/dataloader.py',
> line 279, in iter slash
> return slash _MultiProcessingDataLoaderIter(self) slash
> File
> '/opt/conda/lib/python3.7/site-packages/torch/utils/data/dataloader.py',
> line 719, in init slash
> w.start() slash
> File '/opt/conda/lib/python3.7/multiprocessing/process.py', line 112,
> in start slash
> self. slash _popen = self. slash _Popen(self) slash
> File '/opt/conda/lib/python3.7/multiprocessing/context.py', line 223,
> in slash _Popen slash
> return slash _default_context.get_context().Process. slash _Popen(process_obj) slash
> File '/opt/conda/lib/python3.7/multiprocessing/context.py', line 277,
> in slash _Popen slash
> return Popen(process_obj) slash
> File '/opt/conda/lib/python3.7/multiprocessing/popen_fork.py', line
> 20, in init slash
> self. slash _launch(process_obj) slash
> File '/opt/conda/lib/python3.7/multiprocessing/popen_fork.py', line
> 70, in slash launch slash
> self.pid = os.fork() slash
> OSError: slash [Errno 12 slash ] Cannot allocate memory

Here is the output of my config.json file

> { slash
> 'model': 'Tacotron2', slash
> 'run_name': 'ljspeech', slash
> 'run_description': 'tacotron2',
>
> // AUDIO PARAMETERS
> 'audio':{
> // stft parameters
> 'num_freq': 513, // number of stft frequency levels. Size of the linear spectogram frame.
> 'win_length': 1024, // stft window length in ms.
> 'hop_length': 256, // stft window hop-lengh in ms.
> 'frame_length_ms': null, // stft window length in ms.If null, 'win_length' is used.
> 'frame_shift_ms': null, // stft window hop-lengh in ms. If null, 'hop_length' is used.
>
> // Audio processing parameters
> 'sample_rate': 22050, // DATASET-RELATED: wav sample-rate. If different than the original data, it is resampled.
> 'preemphasis': 0.0, // pre-emphasis to reduce spec noise and make it more structured. If 0.0, no -pre-emphasis.
> 'ref_level_db': 20, // reference level db, theoretically 20db is the sound of air.
>
> // Silence trimming
> 'do_trim_silence': true,// enable trimming of slience of audio as you load it. LJspeech (false), TWEB (false), Nancy (true)
> 'trim_db': 60, // threshold for timming silence. Set this according to your dataset.
>
> // Griffin-Lim
> 'power': 1.5, // value to sharpen wav signals after GL algorithm.
> 'griffin_lim_iters': 60,// #griffin-lim iterations. 30-60 is a good range. Larger the value, slower the generation.
>
> // MelSpectrogram parameters
> 'num_mels': 80, // size of the mel spec frame.
> 'mel_fmin': 0.0, // minimum freq level for mel-spec. ~50 for male and ~95 for female voices. Tune for dataset!!
> 'mel_fmax': 8000.0, // maximum freq level for mel-spec. Tune for dataset!!
>
> // Normalization parameters
> 'signal_norm': true, // normalize spec values. Mean-Var normalization if 'stats_path' is defined otherwise range normalization defined by the other params.
> 'min_level_db': -100, // lower bound for normalization
> 'symmetric_norm': true, // move normalization to range [-1, 1]
> 'max_norm': 4.0, // scale normalization to range [-max_norm, max_norm] or [0, max_norm]
> 'clip_norm': true, // clip normalized values into the range.
> 'stats_path': null // DO NOT USE WITH MULTI_SPEAKER MODEL. scaler stats file computed by 'compute_statistics.py'. If it is defined, mean-std based notmalization is used and other normalization params are ignored
> },
>
> // VOCABULARY PARAMETERS
> // if custom character set is not defined,
> // default set in symbols.py is used
> // 'characters':{
> // 'pad': '',
> // 'eos': '~',
> // 'bos': '^',
> // 'characters': 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz!'(),-.:;? ',
> // 'punctuations':'!'(),-.:;? ',
> // 'phonemes':'iyɨʉɯuɪʏʊeøɘəɵɤoɛœɜɞʌɔæɐaɶɑɒᵻʘɓǀɗǃʄǂɠǁʛpbtdʈɖcɟkɡqɢʔɴŋɲɳnɱmʙrʀⱱɾɽɸβfvθðszʃʒʂʐçʝxɣχʁħʕhɦɬɮʋɹɻjɰlɭʎʟˈˌːˑʍwɥʜʢʡɕʑɺɧɚ˞ɫ'
> // },
>
> // DISTRIBUTED TRAINING
> 'distributed':{
> 'backend': 'nccl',
> 'url': 'tcp: slash / slash /localhost:54321'
> },
>
> 'reinit_layers': [], // give a list of layer names to restore from the given checkpoint. If not defined, it reloads all heuristically matching layers.
>
> // TRAINING
> 'batch_size': 32, // Batch size for training. Lower values than 32 might cause hard to learn attention. It is overwritten by 'gradual_training'.
> 'eval_batch_size':16,
> 'r': 7, // Number of decoder frames to predict per iteration. Set the initial values if gradual training is enabled.
> 'gradual_training': [[0, 7, 64], [1, 5, 64], [50000, 3, 32], [130000, 2, 32], [290000, 1, 32]], //set gradual training steps [first_step, r, batch_size]. If it is null, gradual training is disabled. For Tacotron, you might need to reduce the 'batch_size' as you proceeed.
> 'loss_masking': true, // enable / disable loss masking against the sequence padding.
> 'ga_alpha': 10.0, // weight for guided attention loss. If > 0, guided attention is enabled.
>
> // VALIDATION
> 'run_eval': true,
> 'test_delay_epochs': 10, //Until attention is aligned, testing only wastes computation time.
> 'test_sentences_file': null, // set a file to load sentences to be used for testing. If it is null then we use default english sentences.
>
> // OPTIMIZER
> 'noam_schedule': false, // use noam warmup and lr schedule.
> 'grad_clip': 1.0, // upper limit for gradients for clipping.
> 'epochs': 1000, // total number of epochs to train.
> 'lr': 0.0001, // Initial learning rate. If Noam decay is active, maximum learning rate.
> 'wd': 0.000001, // Weight decay weight.
> 'warmup_steps': 4000, // Noam decay steps to increase the learning rate from 0 to 'lr'
> 'seq_len_norm': false, // Normalize eash sample loss with its length to alleviate imbalanced datasets. Use it if your dataset is small or has skewed distribution of sequence lengths.
>
> // TACOTRON PRENET
> 'memory_size': -1, // ONLY TACOTRON - size of the memory queue used fro storing last decoder predictions for auto-regression. If < 0, memory queue is disabled and decoder only uses the last prediction frame.
> 'prenet_type': 'original', // 'original' or 'bn'.
> 'prenet_dropout': true, // enable/disable dropout at prenet.
>
> // ATTENTION
> 'attention_type': 'original', // 'original' or 'graves'
> 'attention_heads': 4, // number of attention heads (only for 'graves')
> 'attention_norm': 'sigmoid', // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
> 'windowing': false, // Enables attention windowing. Used only in eval mode.
> 'use_forward_attn': false, // if it uses forward attention. In general, it aligns faster.
> 'forward_attn_mask': false, // Additional masking forcing monotonicity only in eval mode.
> 'transition_agent': false, // enable/disable transition agent of forward attention.
> 'location_attn': true, // enable_disable location sensitive attention. It is enabled for TACOTRON by default.
> 'bidirectional_decoder': false, // use https://arxiv.org/abs/1907.09006. Use it, if attention does not work well with your dataset.
>
> // STOPNET
> 'stopnet': true, // Train stopnet predicting the end of synthesis.
> 'separate_stopnet': true, // Train stopnet seperately if 'stopnet==true'. It prevents stopnet loss to influence the rest of the model. It causes a better model, but it trains SLOWER.
>
> // TENSORBOARD and LOGGING
> 'print_step': 25, // Number of steps to log traning on console.
> 'print_eval': false, // If True, it prints intermediate loss values in evalulation.
> 'save_step': 10000, // Number of training steps expected to save traninpg stats and checkpoints.
> 'checkpoint': true, // If true, it saves checkpoints per 'save_step'
> 'tb_model_param_stats': false, // true, plots param stats per layer on tensorboard. Might be memory consuming, but good for debugging.
>
> // DATA LOADING
> 'text_cleaner': 'phoneme_cleaners',
> 'enable_eos_bos_chars': false, // enable/disable beginning of sentence and end of sentence chars.
> 'num_loader_workers': 4, // number of training data loader processes. Don't set it too big. 4-8 are good values.
> 'num_val_loader_workers': 4, // number of evaluation data loader processes.
> 'batch_group_size': 0, //Number of batches to shuffle after bucketing.
> 'min_seq_len': 6, // DATASET-RELATED: minimum text length to use in training
> 'max_seq_len': 153, // DATASET-RELATED: maximum text length
>
> // PATHS
> 'output_path': '../LJSpeech/',
>
> // PHONEMES
> 'phoneme_cache_path': '../TTS/mozilla_us_phonemes_3', // phoneme computation is slow, therefore, it caches results in the given folder.
> 'use_phonemes': true, // use phonemes instead of raw characters. It is suggested for better pronounciation.
> 'phoneme_language': 'it', // depending on your target language, pick one from https://github.com/bootphon/phonemizer#languages
>
> // MULTI-SPEAKER and GST
> 'use_speaker_embedding': false, // use speaker embedding to enable multi-speaker learning.
> 'style_wav_for_test': null, // path to style wav file to be used in TacotronGST inference.
> 'use_gst': false, // TACOTRON ONLY: use global style tokens
>
> // DATASETS
> 'datasets': // List of datasets. They all merged and they get different speaker_ids.
> [
> {
> 'name': 'ljspeech',
> 'path': '../LJSpeech-1.1/',
> 'meta_file_train': 'metadata.csv',
> 'meta_file_val': null
> }
> ]
>
> }

I don't know what happened, can you suggest what is wrong here? slash

[This is an archived TTS discussion thread from discourse.mozilla.org/t/training-fails-after-223-epochs]

[JRMeyer]

>>> erogol
[June 25, 2020, 11:24am]

OSError: slash [Errno 12 slash ] Cannot allocate memory

Your system memory is not enough.

And please do not name tag me. Just imagine if all people here would
name tag me for their problems what would happen to my inbox.

[Archived Post]

[JRMeyer]

>>> Sadam1195
[June 25, 2020, 11:48am]

Okay, got it. Thank you.

> 'num_loader_workers': 2 slash
> 'num_val_loader_workers': 2

I have changed the value to 2 from default 4. Would it help in less
memory operations?

I have also enabled phonemes for Italian, does it perform any memory
operations on CPU? because I have enough memory left on the GPU.

[Archived Post]