Initial commit

config.json

@@ -0,0 +1,76 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
+  "activation_dropout": 0.0,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 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,
+  "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,
+  "final_dropout": 0.0,
+  "gradient_checkpointing": true,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.1,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.1,
+  "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": 0.05,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_attention_heads": 16,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "pad_token_id": 55,
+  "transformers_version": "4.5.0.dev0",
+  "vocab_size": 56
+}

preprocessor_config.json

@@ -0,0 +1,8 @@
+{
+  "do_normalize": true,
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5ea10f6fd23609fd4c28cc96653b2147d12568e092866cdaf7e02bbda40284d
+size 1262163415

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]"}

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": "|"}

trainer_state.json

@@ -0,0 +1,254 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 29.953846153846154,
+  "global_step": 6800,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00023999999999999998,
+      "loss": 6.331,
+      "step": 400
+    },
+    {
+      "epoch": 1.76,
+      "eval_loss": 3.29716157913208,
+      "eval_runtime": 392.7938,
+      "eval_samples_per_second": 8.236,
+      "eval_wer": 0.9908326967150497,
+      "step": 400
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 0.00028573692551505547,
+      "loss": 3.3384,
+      "step": 800
+    },
+    {
+      "epoch": 3.52,
+      "eval_loss": 3.2769880294799805,
+      "eval_runtime": 403.0436,
+      "eval_samples_per_second": 8.026,
+      "eval_wer": 0.9908326967150497,
+      "step": 800
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 0.00026671949286846274,
+      "loss": 2.8767,
+      "step": 1200
+    },
+    {
+      "epoch": 5.29,
+      "eval_loss": 0.987671434879303,
+      "eval_runtime": 414.5553,
+      "eval_samples_per_second": 7.804,
+      "eval_wer": 0.9542438985163444,
+      "step": 1200
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 0.00024770206022187,
+      "loss": 0.7362,
+      "step": 1600
+    },
+    {
+      "epoch": 7.05,
+      "eval_loss": 0.5294700860977173,
+      "eval_runtime": 412.2788,
+      "eval_samples_per_second": 7.847,
+      "eval_wer": 0.7132805275220135,
+      "step": 1600
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 0.0002286846275752773,
+      "loss": 0.4225,
+      "step": 2000
+    },
+    {
+      "epoch": 8.81,
+      "eval_loss": 0.47939780354499817,
+      "eval_runtime": 412.9001,
+      "eval_samples_per_second": 7.835,
+      "eval_wer": 0.6393791966547384,
+      "step": 2000
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 0.0002096671949286846,
+      "loss": 0.2997,
+      "step": 2400
+    },
+    {
+      "epoch": 10.57,
+      "eval_loss": 0.47680211067199707,
+      "eval_runtime": 406.5095,
+      "eval_samples_per_second": 7.958,
+      "eval_wer": 0.5926983233484782,
+      "step": 2400
+    },
+    {
+      "epoch": 12.33,
+      "learning_rate": 0.00019064976228209192,
+      "loss": 0.2349,
+      "step": 2800
+    },
+    {
+      "epoch": 12.33,
+      "eval_loss": 0.46566906571388245,
+      "eval_runtime": 410.6338,
+      "eval_samples_per_second": 7.878,
+      "eval_wer": 0.5762936753648827,
+      "step": 2800
+    },
+    {
+      "epoch": 14.1,
+      "learning_rate": 0.00017163232963549917,
+      "loss": 0.1947,
+      "step": 3200
+    },
+    {
+      "epoch": 14.1,
+      "eval_loss": 0.5099472999572754,
+      "eval_runtime": 412.6367,
+      "eval_samples_per_second": 7.84,
+      "eval_wer": 0.5637087370833501,
+      "step": 3200
+    },
+    {
+      "epoch": 15.86,
+      "learning_rate": 0.00015261489698890646,
+      "loss": 0.1652,
+      "step": 3600
+    },
+    {
+      "epoch": 15.86,
+      "eval_loss": 0.4866821765899658,
+      "eval_runtime": 413.7594,
+      "eval_samples_per_second": 7.819,
+      "eval_wer": 0.5551445458566202,
+      "step": 3600
+    },
+    {
+      "epoch": 17.62,
+      "learning_rate": 0.0001335974643423138,
+      "loss": 0.1397,
+      "step": 4000
+    },
+    {
+      "epoch": 17.62,
+      "eval_loss": 0.510926365852356,
+      "eval_runtime": 416.443,
+      "eval_samples_per_second": 7.768,
+      "eval_wer": 0.5460576575127659,
+      "step": 4000
+    },
+    {
+      "epoch": 19.38,
+      "learning_rate": 0.00011458003169572107,
+      "loss": 0.1251,
+      "step": 4400
+    },
+    {
+      "epoch": 19.38,
+      "eval_loss": 0.5401586294174194,
+      "eval_runtime": 410.6668,
+      "eval_samples_per_second": 7.877,
+      "eval_wer": 0.5362470346990471,
+      "step": 4400
+    },
+    {
+      "epoch": 21.15,
+      "learning_rate": 9.556259904912835e-05,
+      "loss": 0.1105,
+      "step": 4800
+    },
+    {
+      "epoch": 21.15,
+      "eval_loss": 0.5472865104675293,
+      "eval_runtime": 415.7439,
+      "eval_samples_per_second": 7.781,
+      "eval_wer": 0.530416951469583,
+      "step": 4800
+    },
+    {
+      "epoch": 22.91,
+      "learning_rate": 7.654516640253565e-05,
+      "loss": 0.0932,
+      "step": 5200
+    },
+    {
+      "epoch": 22.91,
+      "eval_loss": 0.5360305309295654,
+      "eval_runtime": 420.1088,
+      "eval_samples_per_second": 7.7,
+      "eval_wer": 0.5289694825298541,
+      "step": 5200
+    },
+    {
+      "epoch": 24.67,
+      "learning_rate": 5.752773375594294e-05,
+      "loss": 0.0893,
+      "step": 5600
+    },
+    {
+      "epoch": 24.67,
+      "eval_loss": 0.5389866232872009,
+      "eval_runtime": 429.4038,
+      "eval_samples_per_second": 7.534,
+      "eval_wer": 0.5169072413654457,
+      "step": 5600
+    },
+    {
+      "epoch": 26.43,
+      "learning_rate": 3.851030110935023e-05,
+      "loss": 0.0863,
+      "step": 6000
+    },
+    {
+      "epoch": 26.43,
+      "eval_loss": 0.5444660186767578,
+      "eval_runtime": 423.7501,
+      "eval_samples_per_second": 7.634,
+      "eval_wer": 0.5146958304853042,
+      "step": 6000
+    },
+    {
+      "epoch": 28.19,
+      "learning_rate": 1.9492868462757525e-05,
+      "loss": 0.0778,
+      "step": 6400
+    },
+    {
+      "epoch": 28.19,
+      "eval_loss": 0.5570007562637329,
+      "eval_runtime": 420.4258,
+      "eval_samples_per_second": 7.695,
+      "eval_wer": 0.5113988179003659,
+      "step": 6400
+    },
+    {
+      "epoch": 29.95,
+      "learning_rate": 4.754358161648177e-07,
+      "loss": 0.0732,
+      "step": 6800
+    },
+    {
+      "epoch": 29.95,
+      "eval_loss": 0.557680606842041,
+      "eval_runtime": 427.3219,
+      "eval_samples_per_second": 7.57,
+      "eval_wer": 0.5118411000763942,
+      "step": 6800
+    }
+  ],
+  "max_steps": 6810,
+  "num_train_epochs": 30,
+  "total_flos": 3.375783462381318e+19,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2fa9c7c05a8a24aecbe728ed8d960a674e19cd55c296c7de8bb30b25d9fe68c
+size 2287

vocab.json

@@ -0,0 +1 @@
+{"з": 0, "щ": 2, "x": 3, "ю": 4, "ї": 5, "л": 6, "ж": 7, "с": 8, "—": 9, "ш": 10, "m": 11, "ц": 12, "n": 13, "l": 14, "–": 15, "о": 16, "»": 17, "y": 18, "ч": 19, "и": 20, "a": 21, "м": 22, "o": 23, "i": 24, "c": 25, "п": 26, "«": 27, "p": 28, "…": 29, "ґ": 30, "u": 31, "у": 32, "і": 33, "'": 34, "e": 35, "ф": 36, "ь": 37, "а": 38, "н": 39, "р": 40, "`": 41, "г": 42, "в": 43, "к": 44, "е": 45, "т": 46, "д": 47, "х": 48, "’": 49, "б": 50, "я": 51, "й": 52, "є": 53, "|": 1, "[UNK]": 54, "[PAD]": 55}