Upload 10 files

.gitattributes

@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 Model-ta.nemo filter=lfs diff=lfs merge=lfs -text
+checkpoints/ASR-Model-Language-ta.nemo filter=lfs diff=lfs merge=lfs -text

checkpoints/ASR-Model-Language-ta--val_wer=0.3509-epoch=19.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3bd3f517957625e4f6215303565cbad8d61b4047303d5b870498dbc32f47a68
+size 154197962

checkpoints/ASR-Model-Language-ta--val_wer=0.3509-epoch=20-last.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3ef97644b2b931b104bab71fb1d952eeb118bedb21c48535e9c2749465ac325
+size 154197962

checkpoints/ASR-Model-Language-ta--val_wer=0.3864-epoch=9.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:564d0ebe8855b2ac1190654db705ac315cce0d85e2cc3c2e1d3f8926786a78a6
+size 154197387

checkpoints/ASR-Model-Language-ta.nemo

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14a7ebfb3d783169838f404029db671f15e23cb086d7ee070956cf361ad377d1
+size 147036160

cmd-args.log

@@ -0,0 +1 @@
+/usr/local/lib/python3.10/dist-packages/colab_kernel_launcher.py -f /root/.local/share/jupyter/runtime/kernel-af07951b-f896-4671-afe4-af552f6f2ecd.json

events.out.tfevents.1708036461.0e5e24679ce6.2432.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:610468695b36c3a80ee06f6809baaed1d2e65256a37bd21d4d52de7234482c22
+size 800584

hparams.yaml

@@ -0,0 +1,620 @@
+cfg:
+  sample_rate: 16000
+  train_ds:
+    manifest_filepath: /content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/train/train_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json,/content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/valid/valid_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json
+    sample_rate: 16000
+    batch_size: 16
+    trim_silence: true
+    max_duration: 16.7
+    shuffle: true
+    is_tarred: false
+    tarred_audio_filepaths: null
+    num_workers: 8
+    pin_memory: true
+    use_start_end_token: true
+  validation_ds:
+    manifest_filepath: /content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/test/test_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json
+    sample_rate: 16000
+    batch_size: 8
+    shuffle: false
+    num_workers: 8
+    pin_memory: true
+    use_start_end_token: true
+    trim_silence: true
+  test_ds:
+    manifest_filepath: /content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/test/test_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json
+    sample_rate: 16000
+    batch_size: 8
+    shuffle: false
+    num_workers: 8
+    pin_memory: true
+    use_start_end_token: true
+    trim_silence: true
+  model_defaults:
+    repeat: 5
+    dropout: 0.0
+    separable: true
+    se: true
+    se_context_size: -1
+  tokenizer:
+    dir: tokenizers/ta/tokenizer_spe_bpe_v174/
+    type: bpe
+    model_path: nemo:288fd8b4d3c64e75ac7a6b32b0f3586d_tokenizer.model
+    vocab_path: nemo:6698e9428f25429e8ab2c5238438d52f_vocab.txt
+    spe_tokenizer_vocab: nemo:344d9da9a52049caad7742ae1d994d19_tokenizer.vocab
+  preprocessor:
+    _target_: nemo.collections.asr.modules.AudioToMelSpectrogramPreprocessor
+    sample_rate: 16000
+    normalize: per_feature
+    window_size: 0.025
+    window_stride: 0.01
+    window: hann
+    features: 80
+    n_fft: 512
+    frame_splicing: 1
+    dither: 1.0e-05
+    pad_to: 16
+    stft_conv: false
+  spec_augment:
+    _target_: nemo.collections.asr.modules.SpectrogramAugmentation
+    freq_masks: 2
+    time_masks: 10
+    freq_width: 25
+    time_width: 0.05
+  encoder:
+    _target_: nemo.collections.asr.modules.ConvASREncoder
+    feat_in: 80
+    activation: relu
+    conv_mask: true
+    jasper:
+    - filters: 512
+      repeat: 1
+      kernel:
+      - 5
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: false
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 11
+      stride:
+      - 2
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+      stride_last: true
+      residual_mode: stride_add
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 13
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 15
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 17
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 19
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 21
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 13
+      stride:
+      - 2
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+      stride_last: true
+      residual_mode: stride_add
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 15
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 17
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 19
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 21
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 23
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 25
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 25
+      stride:
+      - 2
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+      stride_last: true
+      residual_mode: stride_add
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 27
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 29
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 31
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 33
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 35
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 37
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 512
+      repeat: 5
+      kernel:
+      - 39
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: true
+      separable: true
+      se: true
+      se_context_size: -1
+    - filters: 640
+      repeat: 1
+      kernel:
+      - 41
+      stride:
+      - 1
+      dilation:
+      - 1
+      dropout: 0.0
+      residual: false
+      separable: true
+      se: true
+      se_context_size: -1
+  decoder:
+    _target_: nemo.collections.asr.modules.ConvASRDecoder
+    feat_in: 640
+    num_classes: 174
+    vocabulary:
+    - <unk>
+    - ்க
+    - ம்
+    - ▁ப
+    - க்க
+    - ்த
+    - ன்
+    - ்ட
+    - ▁வ
+    - ங்க
+    - ரு
+    - ▁இ
+    - ▁க
+    - ▁அ
+    - ▁எ
+    - க்கு
+    - ▁ச
+    - ல்
+    - ந்த
+    - ட்ட
+    - ப்
+    - ▁ந
+    - த்த
+    - து
+    - ப்ப
+    - ▁ம
+    - ல்ல
+    - ▁த
+    - ்ச
+    - ன்ன
+    - ▁இரு
+    - டி
+    - டு
+    - ▁போ
+    - ும்
+    - ந்து
+    - ட்டு
+    - ான்
+    - ாங்க
+    - ச்ச
+    - ிய
+    - ண்
+    - மா
+    - த்து
+    - ▁வந்து
+    - ர்
+    - ▁பா
+    - ண்ண
+    - ▁ஒ
+    - ல்லா
+    - ண்ட
+    - ▁ஆ
+    - ம்ப
+    - ேன்
+    - னு
+    - க்கா
+    - னா
+    - ைய
+    - ▁மா
+    - ▁இருக்கு
+    - ▁கொ
+    - ஸ்
+    - ரி
+    - ▁என்ன
+    - ▁சொ
+    - ▁சா
+    - ச்சு
+    - ள்
+    - ▁ர
+    - ▁பண்ண
+    - ோம்
+    - லா
+    - ▁அப்ப
+    - ட்
+    - ஞ்ச
+    - ▁கா
+    - யி
+    - ய்
+    - ▁எங்க
+    - ▁ஏ
+    - ▁நா
+    - ▁ஒரு
+    - ▁அவ
+    - ீங்க
+    - ியா
+    - ▁அது
+    - ▁எல்லா
+    - ▁கு
+    - தி
+    - ▁இல்ல
+    - ▁வெ
+    - ▁வே
+    - ▁தான்
+    - யா
+    - ▁பே
+    - றது
+    - ▁செ
+    - ுக்கு
+    - ▁இருக்க
+    - ண்டு
+    - ில
+    - ▁பி
+    - ▁ட
+    - ிரு
+    - ளா
+    - ்
+    - ▁
+    - ு
+    - க
+    - ா
+    - த
+    - ட
+    - ப
+    - ம
+    - ி
+    - ன
+    - ர
+    - ல
+    - வ
+    - ச
+    - ந
+    - ங
+    - ய
+    - ே
+    - ோ
+    - ண
+    - இ
+    - .
+    - அ
+    - எ
+    - ள
+    - ை
+    - ெ
+    - ற
+    - ொ
+    - ீ
+    - ஸ
+    - ூ
+    - ஒ
+    - ஆ
+    - ழ
+    - ஞ
+    - ஏ
+    - ஷ
+    - ஜ
+    - ','
+    - உ
+    - ஊ
+    - ஓ
+    - ஃ
+    - ஹ
+    - ஐ
+    - ௌ
+    - ஈ
+    - '!'
+    - '5'
+    - '2'
+    - '0'
+    - '1'
+    - _
+    - '3'
+    - '6'
+    - ஂ
+    - g
+    - k
+    - m
+    - p
+    - s
+    - '௫'
+    - '7'
+    - '8'
+    - '9'
+    - t
+    - '௯'
+  optim:
+    name: novograd
+    lr: 0.025
+    betas:
+    - 0.8
+    - 0.25
+    weight_decay: 0.001
+    sched:
+      name: CosineAnnealing
+      warmup_steps: null
+      warmup_ratio: 0.1
+      min_lr: 1.0e-09
+      last_epoch: -1
+  target: nemo.collections.asr.models.ctc_bpe_models.EncDecCTCModelBPE
+  nemo_version: 1.22.0
+  decoding:
+    strategy: greedy
+    preserve_alignments: null
+    compute_timestamps: null
+    word_seperator: ' '
+    ctc_timestamp_type: all
+    batch_dim_index: 0
+    greedy:
+      preserve_alignments: false
+      compute_timestamps: false
+      preserve_frame_confidence: false
+      confidence_method_cfg:
+        name: entropy
+        entropy_type: tsallis
+        alpha: 0.33
+        entropy_norm: exp
+        temperature: DEPRECATED
+    beam:
+      beam_size: 4
+      search_type: default
+      preserve_alignments: false
+      compute_timestamps: false
+      return_best_hypothesis: true
+      beam_alpha: 1.0
+      beam_beta: 0.0
+      kenlm_path: null
+      flashlight_cfg:
+        lexicon_path: null
+        boost_path: null
+        beam_size_token: 16
+        beam_threshold: 20.0
+        unk_weight: -.inf
+        sil_weight: 0.0
+      pyctcdecode_cfg:
+        beam_prune_logp: -10.0
+        token_min_logp: -5.0
+        prune_history: false
+        hotwords: null
+        hotword_weight: 10.0
+    confidence_cfg:
+      preserve_frame_confidence: false
+      preserve_token_confidence: false
+      preserve_word_confidence: false
+      exclude_blank: true
+      aggregation: min
+      method_cfg:
+        name: entropy
+        entropy_type: tsallis
+        alpha: 0.33
+        entropy_norm: exp
+        temperature: DEPRECATED
+    temperature: 1.0

lightning_logs.txt

@@ -0,0 +1,24 @@
+GPU available: True (cuda), used: True
+TPU available: False, using: 0 TPU cores
+IPU available: False, using: 0 IPUs
+HPU available: False, using: 0 HPUs
+LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]
+
+  | Name              | Type                              | Params
+------------------------------------------------------------------------
+0 | preprocessor      | AudioToMelSpectrogramPreprocessor | 0     
+1 | encoder           | ConvASREncoder                    | 36.3 M
+2 | decoder           | ConvASRDecoder                    | 112 K 
+3 | loss              | CTCLoss                           | 0     
+4 | spec_augmentation | SpectrogramAugmentation           | 0     
+5 | wer               | WER                               | 0     
+------------------------------------------------------------------------
+1.8 M     Trainable params
+34.7 M    Non-trainable params
+36.4 M    Total params
+145.798   Total estimated model params size (MB)
+Epoch 9, global step 5070: 'val_wer' reached 0.38640 (best 0.38640), saving model to '/content/experiments/lang-ta/ASR-Model-Language-ta/2024-02-15_22-34-08/checkpoints/ASR-Model-Language-ta--val_wer=0.3864-epoch=9.ckpt' as top 3
+Epoch 19, global step 10140: 'val_wer' reached 0.35093 (best 0.35093), saving model to '/content/experiments/lang-ta/ASR-Model-Language-ta/2024-02-15_22-34-08/checkpoints/ASR-Model-Language-ta--val_wer=0.3509-epoch=19.ckpt' as top 3
+`Trainer.fit` stopped: `max_epochs=20` reached.
+Restoring states from the checkpoint path at /content/experiments/lang-ta/ASR-Model-Language-ta/2024-02-15_22-34-08/checkpoints/ASR-Model-Language-ta--val_wer=0.3509-epoch=19.ckpt
+Restored all states from the checkpoint at /content/experiments/lang-ta/ASR-Model-Language-ta/2024-02-15_22-34-08/checkpoints/ASR-Model-Language-ta--val_wer=0.3509-epoch=19.ckpt

nemo_error_log.txt

@@ -0,0 +1,42 @@
+[NeMo W 2024-02-15 22:31:14 modelPT:165] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.
+    Train config : 
+    manifest_filepath: /content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/train/train_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json,/content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/valid/valid_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json
+    sample_rate: 16000
+    batch_size: 16
+    trim_silence: true
+    max_duration: 16.7
+    shuffle: true
+    is_tarred: false
+    tarred_audio_filepaths: null
+    num_workers: 8
+    pin_memory: true
+    use_start_end_token: true
+    
+[NeMo W 2024-02-15 22:31:14 modelPT:172] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s). 
+    Validation config : 
+    manifest_filepath: /content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/test/test_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json
+    sample_rate: 16000
+    batch_size: 8
+    shuffle: false
+    num_workers: 8
+    pin_memory: true
+    use_start_end_token: true
+    trim_silence: true
+    
+[NeMo W 2024-02-15 22:31:14 modelPT:178] Please call the ModelPT.setup_test_data() or ModelPT.setup_multiple_test_data() method and provide a valid configuration file to setup the test data loader(s).
+    Test config : 
+    manifest_filepath: /content/datasets/ta/yaygomii/Tamil-Speech-Dialect-Corpus-Shuffled-Split/test/test_yaygomii_Tamil-Speech-Dialect-Corpus-Shuffled-Split_manifest_processed.json
+    sample_rate: 16000
+    batch_size: 8
+    shuffle: false
+    num_workers: 8
+    pin_memory: true
+    use_start_end_token: true
+    trim_silence: true
+    
+[NeMo W 2024-02-15 22:32:34 nemo_logging:349] /usr/local/lib/python3.10/dist-packages/torch/utils/data/dataloader.py:557: UserWarning: This DataLoader will create 8 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.
+      warnings.warn(_create_warning_msg(
+    
+[NeMo W 2024-02-15 22:34:21 nemo_logging:349] /usr/local/lib/python3.10/dist-packages/torch/utils/data/dataloader.py:557: UserWarning: This DataLoader will create 8 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.
+      warnings.warn(_create_warning_msg(
+