First model version

added_tokens.json

@@ -0,0 +1 @@
+{"<s>": 63, "</s>": 64}

all_results.json

@@ -0,0 +1,20 @@
+{
+    "epoch": 15.0,
+    "eval_loss": 0.5424522757530212,
+    "eval_runtime": 7.8831,
+    "eval_samples": 264,
+    "eval_samples_per_second": 33.49,
+    "eval_steps_per_second": 4.186,
+    "eval_wer": 0.6865671641791045,
+    "test_loss": 0.5638678669929504,
+    "test_runtime": 8.612,
+    "test_samples": 264,
+    "test_samples_per_second": 30.655,
+    "test_steps_per_second": 3.832,
+    "test_wer": 0.6888412017167382,
+    "train_loss": 1.7177187274751209,
+    "train_runtime": 3292.1945,
+    "train_samples": 4465,
+    "train_samples_per_second": 20.344,
+    "train_steps_per_second": 0.638
+}

config.json

@@ -0,0 +1,86 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
+  "activation_dropout": 0.0,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "classifier_proj_size": 256,
+  "codevector_dim": 768,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": false,
+  "diversity_loss_weight": 0.1,
+  "do_stable_layer_norm": true,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.0,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.0,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.0,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "mask_feature_length": 10,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 9,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.01,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_attention_heads": 16,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "num_negatives": 100,
+  "pad_token_id": 62,
+  "proj_codevector_dim": 768,
+  "torch_dtype": "float32",
+  "transformers_version": "4.12.0.dev0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 65
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 15.0,
+    "eval_loss": 0.5424522757530212,
+    "eval_runtime": 7.8831,
+    "eval_samples": 264,
+    "eval_samples_per_second": 33.49,
+    "eval_steps_per_second": 4.186,
+    "eval_wer": 0.6865671641791045
+}

preprocessor_config.json

@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9fbbaf8c7fae1806960bf43dcdec04415249ab919505a5a60bd8c4089592c6e
+size 1262195308

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}

test_results.json

@@ -0,0 +1,8 @@
+{
+    "test_loss": 0.5638678669929504,
+    "test_runtime": 8.612,
+    "test_samples": 264,
+    "test_samples_per_second": 30.655,
+    "test_steps_per_second": 3.832,
+    "test_wer": 0.6888412017167382
+}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|", "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "./wav2vec2-chr_voice_private-ep15-lr0.0003-repeat0-mask0.01", "tokenizer_class": "Wav2Vec2CTCTokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 15.0,
+    "train_loss": 1.7177187274751209,
+    "train_runtime": 3292.1945,
+    "train_samples": 4465,
+    "train_samples_per_second": 20.344,
+    "train_steps_per_second": 0.638
+}

trainer_state.json

@@ -0,0 +1,427 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 15.0,
+  "global_step": 2100,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.36,
+      "eval_loss": 4.132408142089844,
+      "eval_runtime": 7.7892,
+      "eval_samples_per_second": 33.893,
+      "eval_steps_per_second": 4.237,
+      "eval_wer": 1.0,
+      "step": 50
+    },
+    {
+      "epoch": 0.71,
+      "eval_loss": 3.357618570327759,
+      "eval_runtime": 7.7699,
+      "eval_samples_per_second": 33.977,
+      "eval_steps_per_second": 4.247,
+      "eval_wer": 1.0,
+      "step": 100
+    },
+    {
+      "epoch": 1.07,
+      "eval_loss": 3.0935420989990234,
+      "eval_runtime": 8.3088,
+      "eval_samples_per_second": 31.773,
+      "eval_steps_per_second": 3.972,
+      "eval_wer": 1.0,
+      "step": 150
+    },
+    {
+      "epoch": 1.43,
+      "eval_loss": 3.000765085220337,
+      "eval_runtime": 7.8476,
+      "eval_samples_per_second": 33.641,
+      "eval_steps_per_second": 4.205,
+      "eval_wer": 1.2494669509594882,
+      "step": 200
+    },
+    {
+      "epoch": 1.79,
+      "eval_loss": 2.8665499687194824,
+      "eval_runtime": 7.9301,
+      "eval_samples_per_second": 33.291,
+      "eval_steps_per_second": 4.161,
+      "eval_wer": 1.0874200426439233,
+      "step": 250
+    },
+    {
+      "epoch": 2.14,
+      "eval_loss": 2.7977161407470703,
+      "eval_runtime": 7.8406,
+      "eval_samples_per_second": 33.671,
+      "eval_steps_per_second": 4.209,
+      "eval_wer": 1.537313432835821,
+      "step": 300
+    },
+    {
+      "epoch": 2.5,
+      "eval_loss": 2.618842601776123,
+      "eval_runtime": 8.0288,
+      "eval_samples_per_second": 32.882,
+      "eval_steps_per_second": 4.11,
+      "eval_wer": 1.2345415778251598,
+      "step": 350
+    },
+    {
+      "epoch": 2.86,
+      "eval_loss": 2.3096766471862793,
+      "eval_runtime": 8.0346,
+      "eval_samples_per_second": 32.858,
+      "eval_steps_per_second": 4.107,
+      "eval_wer": 1.1279317697228144,
+      "step": 400
+    },
+    {
+      "epoch": 3.21,
+      "eval_loss": 1.8073012828826904,
+      "eval_runtime": 7.9049,
+      "eval_samples_per_second": 33.397,
+      "eval_steps_per_second": 4.175,
+      "eval_wer": 1.251599147121535,
+      "step": 450
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 0.0002409,
+      "loss": 3.5589,
+      "step": 500
+    },
+    {
+      "epoch": 3.57,
+      "eval_loss": 1.3744713068008423,
+      "eval_runtime": 7.9139,
+      "eval_samples_per_second": 33.359,
+      "eval_steps_per_second": 4.17,
+      "eval_wer": 1.0895522388059702,
+      "step": 500
+    },
+    {
+      "epoch": 3.93,
+      "eval_loss": 1.1971436738967896,
+      "eval_runtime": 8.4443,
+      "eval_samples_per_second": 31.264,
+      "eval_steps_per_second": 3.908,
+      "eval_wer": 1.2921108742004264,
+      "step": 550
+    },
+    {
+      "epoch": 4.29,
+      "eval_loss": 1.0361448526382446,
+      "eval_runtime": 7.8907,
+      "eval_samples_per_second": 33.457,
+      "eval_steps_per_second": 4.182,
+      "eval_wer": 0.9872068230277186,
+      "step": 600
+    },
+    {
+      "epoch": 4.64,
+      "eval_loss": 1.0113328695297241,
+      "eval_runtime": 8.6808,
+      "eval_samples_per_second": 30.412,
+      "eval_steps_per_second": 3.802,
+      "eval_wer": 1.1556503198294243,
+      "step": 650
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 0.9761010408401489,
+      "eval_runtime": 8.374,
+      "eval_samples_per_second": 31.526,
+      "eval_steps_per_second": 3.941,
+      "eval_wer": 0.9509594882729211,
+      "step": 700
+    },
+    {
+      "epoch": 5.36,
+      "eval_loss": 0.8795022368431091,
+      "eval_runtime": 8.0897,
+      "eval_samples_per_second": 32.634,
+      "eval_steps_per_second": 4.079,
+      "eval_wer": 1.1279317697228144,
+      "step": 750
+    },
+    {
+      "epoch": 5.71,
+      "eval_loss": 0.8115519881248474,
+      "eval_runtime": 7.9853,
+      "eval_samples_per_second": 33.061,
+      "eval_steps_per_second": 4.133,
+      "eval_wer": 0.8869936034115139,
+      "step": 800
+    },
+    {
+      "epoch": 6.07,
+      "eval_loss": 0.7683095932006836,
+      "eval_runtime": 8.0654,
+      "eval_samples_per_second": 32.732,
+      "eval_steps_per_second": 4.092,
+      "eval_wer": 0.9275053304904051,
+      "step": 850
+    },
+    {
+      "epoch": 6.43,
+      "eval_loss": 0.7249290943145752,
+      "eval_runtime": 7.8449,
+      "eval_samples_per_second": 33.652,
+      "eval_steps_per_second": 4.207,
+      "eval_wer": 1.0255863539445629,
+      "step": 900
+    },
+    {
+      "epoch": 6.79,
+      "eval_loss": 0.7122178077697754,
+      "eval_runtime": 8.1435,
+      "eval_samples_per_second": 32.419,
+      "eval_steps_per_second": 4.052,
+      "eval_wer": 0.9211087420042644,
+      "step": 950
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 0.00016634999999999998,
+      "loss": 1.5095,
+      "step": 1000
+    },
+    {
+      "epoch": 7.14,
+      "eval_loss": 0.7041318416595459,
+      "eval_runtime": 9.0145,
+      "eval_samples_per_second": 29.286,
+      "eval_steps_per_second": 3.661,
+      "eval_wer": 1.0319829424307037,
+      "step": 1000
+    },
+    {
+      "epoch": 7.5,
+      "eval_loss": 0.678531289100647,
+      "eval_runtime": 8.7862,
+      "eval_samples_per_second": 30.047,
+      "eval_steps_per_second": 3.756,
+      "eval_wer": 0.8699360341151386,
+      "step": 1050
+    },
+    {
+      "epoch": 7.86,
+      "eval_loss": 0.7056036591529846,
+      "eval_runtime": 8.0033,
+      "eval_samples_per_second": 32.986,
+      "eval_steps_per_second": 4.123,
+      "eval_wer": 0.9680170575692963,
+      "step": 1100
+    },
+    {
+      "epoch": 8.21,
+      "eval_loss": 0.6487303972244263,
+      "eval_runtime": 8.7104,
+      "eval_samples_per_second": 30.309,
+      "eval_steps_per_second": 3.789,
+      "eval_wer": 0.8550106609808102,
+      "step": 1150
+    },
+    {
+      "epoch": 8.57,
+      "eval_loss": 0.5972908139228821,
+      "eval_runtime": 7.951,
+      "eval_samples_per_second": 33.204,
+      "eval_steps_per_second": 4.15,
+      "eval_wer": 0.7889125799573561,
+      "step": 1200
+    },
+    {
+      "epoch": 8.93,
+      "eval_loss": 0.5955255627632141,
+      "eval_runtime": 7.9427,
+      "eval_samples_per_second": 33.238,
+      "eval_steps_per_second": 4.155,
+      "eval_wer": 0.8443496801705757,
+      "step": 1250
+    },
+    {
+      "epoch": 9.29,
+      "eval_loss": 0.5822768211364746,
+      "eval_runtime": 7.8596,
+      "eval_samples_per_second": 33.59,
+      "eval_steps_per_second": 4.199,
+      "eval_wer": 0.8017057569296375,
+      "step": 1300
+    },
+    {
+      "epoch": 9.64,
+      "eval_loss": 0.5886873006820679,
+      "eval_runtime": 7.8662,
+      "eval_samples_per_second": 33.561,
+      "eval_steps_per_second": 4.195,
+      "eval_wer": 0.7569296375266524,
+      "step": 1350
+    },
+    {
+      "epoch": 10.0,
+      "eval_loss": 0.5869713425636292,
+      "eval_runtime": 7.8909,
+      "eval_samples_per_second": 33.456,
+      "eval_steps_per_second": 4.182,
+      "eval_wer": 0.7569296375266524,
+      "step": 1400
+    },
+    {
+      "epoch": 10.36,
+      "eval_loss": 0.5845889449119568,
+      "eval_runtime": 7.8445,
+      "eval_samples_per_second": 33.654,
+      "eval_steps_per_second": 4.207,
+      "eval_wer": 0.7484008528784648,
+      "step": 1450
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 9.18e-05,
+      "loss": 1.1157,
+      "step": 1500
+    },
+    {
+      "epoch": 10.71,
+      "eval_loss": 0.5864734053611755,
+      "eval_runtime": 8.0229,
+      "eval_samples_per_second": 32.906,
+      "eval_steps_per_second": 4.113,
+      "eval_wer": 0.7547974413646056,
+      "step": 1500
+    },
+    {
+      "epoch": 11.07,
+      "eval_loss": 0.5586370825767517,
+      "eval_runtime": 7.8673,
+      "eval_samples_per_second": 33.557,
+      "eval_steps_per_second": 4.195,
+      "eval_wer": 0.7334754797441365,
+      "step": 1550
+    },
+    {
+      "epoch": 11.43,
+      "eval_loss": 0.5573432445526123,
+      "eval_runtime": 7.8679,
+      "eval_samples_per_second": 33.554,
+      "eval_steps_per_second": 4.194,
+      "eval_wer": 0.744136460554371,
+      "step": 1600
+    },
+    {
+      "epoch": 11.79,
+      "eval_loss": 0.5594019889831543,
+      "eval_runtime": 7.9618,
+      "eval_samples_per_second": 33.158,
+      "eval_steps_per_second": 4.145,
+      "eval_wer": 0.7292110874200426,
+      "step": 1650
+    },
+    {
+      "epoch": 12.14,
+      "eval_loss": 0.5614868998527527,
+      "eval_runtime": 7.8272,
+      "eval_samples_per_second": 33.729,
+      "eval_steps_per_second": 4.216,
+      "eval_wer": 0.7569296375266524,
+      "step": 1700
+    },
+    {
+      "epoch": 12.5,
+      "eval_loss": 0.5569693446159363,
+      "eval_runtime": 8.0892,
+      "eval_samples_per_second": 32.636,
+      "eval_steps_per_second": 4.08,
+      "eval_wer": 0.7654584221748401,
+      "step": 1750
+    },
+    {
+      "epoch": 12.86,
+      "eval_loss": 0.5408880710601807,
+      "eval_runtime": 7.8701,
+      "eval_samples_per_second": 33.545,
+      "eval_steps_per_second": 4.193,
+      "eval_wer": 0.7121535181236673,
+      "step": 1800
+    },
+    {
+      "epoch": 13.21,
+      "eval_loss": 0.5358032584190369,
+      "eval_runtime": 7.8977,
+      "eval_samples_per_second": 33.427,
+      "eval_steps_per_second": 4.178,
+      "eval_wer": 0.6652452025586354,
+      "step": 1850
+    },
+    {
+      "epoch": 13.57,
+      "eval_loss": 0.5394359827041626,
+      "eval_runtime": 7.8219,
+      "eval_samples_per_second": 33.751,
+      "eval_steps_per_second": 4.219,
+      "eval_wer": 0.6823027718550106,
+      "step": 1900
+    },
+    {
+      "epoch": 13.93,
+      "eval_loss": 0.5434439778327942,
+      "eval_runtime": 7.8516,
+      "eval_samples_per_second": 33.624,
+      "eval_steps_per_second": 4.203,
+      "eval_wer": 0.6993603411513859,
+      "step": 1950
+    },
+    {
+      "epoch": 14.29,
+      "learning_rate": 1.74e-05,
+      "loss": 0.8658,
+      "step": 2000
+    },
+    {
+      "epoch": 14.29,
+      "eval_loss": 0.5396074056625366,
+      "eval_runtime": 7.9269,
+      "eval_samples_per_second": 33.304,
+      "eval_steps_per_second": 4.163,
+      "eval_wer": 0.6823027718550106,
+      "step": 2000
+    },
+    {
+      "epoch": 14.64,
+      "eval_loss": 0.5431792736053467,
+      "eval_runtime": 7.8451,
+      "eval_samples_per_second": 33.651,
+      "eval_steps_per_second": 4.206,
+      "eval_wer": 0.6780383795309168,
+      "step": 2050
+    },
+    {
+      "epoch": 15.0,
+      "eval_loss": 0.5424522757530212,
+      "eval_runtime": 7.9063,
+      "eval_samples_per_second": 33.391,
+      "eval_steps_per_second": 4.174,
+      "eval_wer": 0.6865671641791045,
+      "step": 2100
+    },
+    {
+      "epoch": 15.0,
+      "step": 2100,
+      "total_flos": 2.38289603930769e+18,
+      "train_loss": 1.7177187274751209,
+      "train_runtime": 3292.1945,
+      "train_samples_per_second": 20.344,
+      "train_steps_per_second": 0.638
+    }
+  ],
+  "max_steps": 2100,
+  "num_train_epochs": 15,
+  "total_flos": 2.38289603930769e+18,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0c99ff9e12aa55a94bfb7bbe4e77587309782cf5256f9afb21a7656a737b0b6
+size 2863

vocab.json

@@ -0,0 +1 @@
+{":": 1, ";": 2, "a": 3, "b": 4, "c": 5, "d": 6, "e": 7, "g": 8, "h": 9, "i": 10, "j": 11, "k": 12, "l": 13, "m": 14, "n": 15, "o": 16, "r": 17, "s": 18, "t": 19, "u": 20, "v": 21, "w": 22, "y": 23, "à": 24, "á": 25, "â": 26, "è": 27, "é": 28, "ê": 29, "ì": 30, "í": 31, "î": 32, "ò": 33, "ó": 34, "ô": 35, "ö": 36, "ù": 37, "ú": 38, "û": 39, "ā": 40, "ē": 41, "ě": 42, "ī": 43, "ō": 44, "ő": 45, "ū": 46, "ű": 47, "ǎ": 48, "ǐ": 49, "ǒ": 50, "ǔ": 51, "ɂ": 52, "̀": 53, "́": 54, "̂": 55, "̄": 56, "̋": 57, "̌": 58, "“": 59, "”": 60, "|": 0, "[UNK]": 61, "[PAD]": 62}