First model version

added_tokens.json

@@ -0,0 +1 @@
+{"<s>": 65, "</s>": 66}

all_results.json

@@ -0,0 +1,20 @@
+{
+    "epoch": 14.99,
+    "eval_loss": 0.6640865206718445,
+    "eval_runtime": 6.6949,
+    "eval_samples": 264,
+    "eval_samples_per_second": 39.433,
+    "eval_steps_per_second": 4.929,
+    "eval_wer": 0.7348178137651822,
+    "test_loss": 0.676352322101593,
+    "test_runtime": 7.2219,
+    "test_samples": 264,
+    "test_samples_per_second": 36.556,
+    "test_steps_per_second": 4.569,
+    "test_wer": 0.7626373626373626,
+    "train_loss": 2.406289642025726,
+    "train_runtime": 1161.5479,
+    "train_samples": 2120,
+    "train_samples_per_second": 27.377,
+    "train_steps_per_second": 0.852
+}

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": 10,
+  "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": 64,
+  "proj_codevector_dim": 768,
+  "torch_dtype": "float32",
+  "transformers_version": "4.12.0.dev0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 67
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 14.99,
+    "eval_loss": 0.6640865206718445,
+    "eval_runtime": 6.6949,
+    "eval_samples": 264,
+    "eval_samples_per_second": 39.433,
+    "eval_steps_per_second": 4.929,
+    "eval_wer": 0.7348178137651822
+}

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:a5ba0a05bd07a692bdb922a68b5b7c678ce452294b2285936c227fbf1ed27617
+size 1262203433

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": 0.676352322101593,
+    "test_runtime": 7.2219,
+    "test_samples": 264,
+    "test_samples_per_second": 36.556,
+    "test_steps_per_second": 4.569,
+    "test_wer": 0.7626373626373626
+}

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-ep15-lr0.0003-repeat0-mask0.01", "tokenizer_class": "Wav2Vec2CTCTokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 14.99,
+    "train_loss": 2.406289642025726,
+    "train_runtime": 1161.5479,
+    "train_samples": 2120,
+    "train_samples_per_second": 27.377,
+    "train_steps_per_second": 0.852
+}

trainer_state.json

@@ -0,0 +1,202 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 14.992481203007518,
+  "global_step": 990,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.75,
+      "eval_loss": 4.109964847564697,
+      "eval_runtime": 6.6506,
+      "eval_samples_per_second": 39.695,
+      "eval_steps_per_second": 4.962,
+      "eval_wer": 1.0,
+      "step": 50
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 3.392470121383667,
+      "eval_runtime": 6.6261,
+      "eval_samples_per_second": 39.842,
+      "eval_steps_per_second": 4.98,
+      "eval_wer": 1.0,
+      "step": 100
+    },
+    {
+      "epoch": 2.27,
+      "eval_loss": 3.1406397819519043,
+      "eval_runtime": 6.687,
+      "eval_samples_per_second": 39.479,
+      "eval_steps_per_second": 4.935,
+      "eval_wer": 1.008097165991903,
+      "step": 150
+    },
+    {
+      "epoch": 3.03,
+      "eval_loss": 3.0505788326263428,
+      "eval_runtime": 6.6386,
+      "eval_samples_per_second": 39.767,
+      "eval_steps_per_second": 4.971,
+      "eval_wer": 1.2834008097165992,
+      "step": 200
+    },
+    {
+      "epoch": 3.78,
+      "eval_loss": 2.8735897541046143,
+      "eval_runtime": 6.5891,
+      "eval_samples_per_second": 40.066,
+      "eval_steps_per_second": 5.008,
+      "eval_wer": 1.0323886639676114,
+      "step": 250
+    },
+    {
+      "epoch": 4.54,
+      "eval_loss": 2.7715253829956055,
+      "eval_runtime": 6.6425,
+      "eval_samples_per_second": 39.744,
+      "eval_steps_per_second": 4.968,
+      "eval_wer": 1.1437246963562753,
+      "step": 300
+    },
+    {
+      "epoch": 5.3,
+      "eval_loss": 2.5352094173431396,
+      "eval_runtime": 6.6247,
+      "eval_samples_per_second": 39.851,
+      "eval_steps_per_second": 4.981,
+      "eval_wer": 1.0101214574898785,
+      "step": 350
+    },
+    {
+      "epoch": 6.06,
+      "eval_loss": 2.0723843574523926,
+      "eval_runtime": 6.6216,
+      "eval_samples_per_second": 39.869,
+      "eval_steps_per_second": 4.984,
+      "eval_wer": 1.0263157894736843,
+      "step": 400
+    },
+    {
+      "epoch": 6.81,
+      "eval_loss": 1.4460173845291138,
+      "eval_runtime": 6.6321,
+      "eval_samples_per_second": 39.806,
+      "eval_steps_per_second": 4.976,
+      "eval_wer": 0.9696356275303644,
+      "step": 450
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 0.00016584269662921346,
+      "loss": 3.4597,
+      "step": 500
+    },
+    {
+      "epoch": 7.57,
+      "eval_loss": 1.1178003549575806,
+      "eval_runtime": 6.6144,
+      "eval_samples_per_second": 39.913,
+      "eval_steps_per_second": 4.989,
+      "eval_wer": 0.8461538461538461,
+      "step": 500
+    },
+    {
+      "epoch": 8.33,
+      "eval_loss": 0.950170636177063,
+      "eval_runtime": 6.8415,
+      "eval_samples_per_second": 38.588,
+      "eval_steps_per_second": 4.823,
+      "eval_wer": 0.8461538461538461,
+      "step": 550
+    },
+    {
+      "epoch": 9.09,
+      "eval_loss": 0.8674038052558899,
+      "eval_runtime": 6.727,
+      "eval_samples_per_second": 39.245,
+      "eval_steps_per_second": 4.906,
+      "eval_wer": 0.7894736842105263,
+      "step": 600
+    },
+    {
+      "epoch": 9.84,
+      "eval_loss": 0.7972543239593506,
+      "eval_runtime": 6.6391,
+      "eval_samples_per_second": 39.765,
+      "eval_steps_per_second": 4.971,
+      "eval_wer": 0.7692307692307693,
+      "step": 650
+    },
+    {
+      "epoch": 10.6,
+      "eval_loss": 0.7665630578994751,
+      "eval_runtime": 6.8502,
+      "eval_samples_per_second": 38.539,
+      "eval_steps_per_second": 4.817,
+      "eval_wer": 0.7753036437246964,
+      "step": 700
+    },
+    {
+      "epoch": 11.36,
+      "eval_loss": 0.7286367416381836,
+      "eval_runtime": 8.3531,
+      "eval_samples_per_second": 31.605,
+      "eval_steps_per_second": 3.951,
+      "eval_wer": 0.7530364372469636,
+      "step": 750
+    },
+    {
+      "epoch": 12.12,
+      "eval_loss": 0.7024735808372498,
+      "eval_runtime": 6.7477,
+      "eval_samples_per_second": 39.125,
+      "eval_steps_per_second": 4.891,
+      "eval_wer": 0.7307692307692307,
+      "step": 800
+    },
+    {
+      "epoch": 12.87,
+      "eval_loss": 0.6833031177520752,
+      "eval_runtime": 6.8878,
+      "eval_samples_per_second": 38.328,
+      "eval_steps_per_second": 4.791,
+      "eval_wer": 0.7348178137651822,
+      "step": 850
+    },
+    {
+      "epoch": 13.63,
+      "eval_loss": 0.669713020324707,
+      "eval_runtime": 6.7425,
+      "eval_samples_per_second": 39.154,
+      "eval_steps_per_second": 4.894,
+      "eval_wer": 0.7226720647773279,
+      "step": 900
+    },
+    {
+      "epoch": 14.39,
+      "eval_loss": 0.6649630069732666,
+      "eval_runtime": 6.8935,
+      "eval_samples_per_second": 38.297,
+      "eval_steps_per_second": 4.787,
+      "eval_wer": 0.7368421052631579,
+      "step": 950
+    },
+    {
+      "epoch": 14.99,
+      "step": 990,
+      "total_flos": 1.166783970335839e+18,
+      "train_loss": 2.406289642025726,
+      "train_runtime": 1161.5479,
+      "train_samples_per_second": 27.377,
+      "train_steps_per_second": 0.852
+    }
+  ],
+  "max_steps": 990,
+  "num_train_epochs": 15,
+  "total_flos": 1.166783970335839e+18,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0d4f9d359f70903c83f79948b43fd7be0a876462537b02758a5bf6e5045f305
+size 2863

vocab.json

@@ -0,0 +1 @@
+{"(": 1, ")": 2, ":": 3, ";": 4, "a": 5, "b": 6, "c": 7, "d": 8, "e": 9, "g": 10, "h": 11, "i": 12, "j": 13, "k": 14, "l": 15, "m": 16, "n": 17, "o": 18, "r": 19, "s": 20, "t": 21, "u": 22, "v": 23, "w": 24, "y": 25, "à": 26, "á": 27, "â": 28, "è": 29, "é": 30, "ê": 31, "ì": 32, "í": 33, "î": 34, "ò": 35, "ó": 36, "ô": 37, "ö": 38, "ù": 39, "ú": 40, "û": 41, "ā": 42, "ē": 43, "ě": 44, "ī": 45, "ō": 46, "ő": 47, "ū": 48, "ű": 49, "ǎ": 50, "ǐ": 51, "ǒ": 52, "ǔ": 53, "ɂ": 54, "̀": 55, "́": 56, "̂": 57, "̄": 58, "̋": 59, "̌": 60, "“": 61, "”": 62, "|": 0, "[UNK]": 63, "[PAD]": 64}