First model version

added_tokens.json

@@ -0,0 +1 @@
+{"<s>": 88, "</s>": 89}

all_results.json

@@ -0,0 +1,20 @@
+{
+    "epoch": 14.99,
+    "eval_loss": 0.1987980157136917,
+    "eval_runtime": 25.4983,
+    "eval_samples": 205,
+    "eval_samples_per_second": 8.04,
+    "eval_steps_per_second": 1.02,
+    "eval_wer": 0.20191649555099248,
+    "test_loss": 0.20211350917816162,
+    "test_runtime": 25.1371,
+    "test_samples": 205,
+    "test_samples_per_second": 8.155,
+    "test_steps_per_second": 1.034,
+    "test_wer": 0.23429781227946367,
+    "train_loss": 2.7169430053312014,
+    "train_runtime": 6457.8468,
+    "train_samples": 1648,
+    "train_samples_per_second": 3.828,
+    "train_steps_per_second": 0.118
+}

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": 87,
+  "proj_codevector_dim": 768,
+  "torch_dtype": "float32",
+  "transformers_version": "4.12.0.dev0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 90
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 14.99,
+    "eval_loss": 0.1987980157136917,
+    "eval_runtime": 25.4983,
+    "eval_samples": 205,
+    "eval_samples_per_second": 8.04,
+    "eval_steps_per_second": 1.02,
+    "eval_wer": 0.20191649555099248
+}

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 sha256:97a23eb47e3921c0edb9e8f72caa62a09bc85173d9e9b263f9d86f8d313a5296
+size 1262297780

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": 0.20211350917816162,
+    "test_runtime": 25.1371,
+    "test_samples": 205,
+    "test_samples_per_second": 8.155,
+    "test_steps_per_second": 1.034,
+    "test_wer": 0.23429781227946367
+}

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

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 14.99,
+    "train_loss": 2.7169430053312014,
+    "train_runtime": 6457.8468,
+    "train_samples": 1648,
+    "train_samples_per_second": 3.828,
+    "train_steps_per_second": 0.118
+}

trainer_state.json

@@ -0,0 +1,166 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 14.990291262135923,
+  "global_step": 765,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.97,
+      "eval_loss": 6.859605312347412,
+      "eval_runtime": 25.516,
+      "eval_samples_per_second": 8.034,
+      "eval_steps_per_second": 1.019,
+      "eval_wer": 1.0,
+      "step": 50
+    },
+    {
+      "epoch": 1.95,
+      "eval_loss": 3.7317371368408203,
+      "eval_runtime": 25.4287,
+      "eval_samples_per_second": 8.062,
+      "eval_steps_per_second": 1.022,
+      "eval_wer": 1.0,
+      "step": 100
+    },
+    {
+      "epoch": 2.93,
+      "eval_loss": 3.4334847927093506,
+      "eval_runtime": 25.371,
+      "eval_samples_per_second": 8.08,
+      "eval_steps_per_second": 1.025,
+      "eval_wer": 1.0,
+      "step": 150
+    },
+    {
+      "epoch": 3.91,
+      "eval_loss": 1.2018941640853882,
+      "eval_runtime": 25.2525,
+      "eval_samples_per_second": 8.118,
+      "eval_steps_per_second": 1.03,
+      "eval_wer": 0.7789185489390829,
+      "step": 200
+    },
+    {
+      "epoch": 4.89,
+      "eval_loss": 0.3254333734512329,
+      "eval_runtime": 25.4191,
+      "eval_samples_per_second": 8.065,
+      "eval_steps_per_second": 1.023,
+      "eval_wer": 0.3333333333333333,
+      "step": 250
+    },
+    {
+      "epoch": 5.87,
+      "eval_loss": 0.2722693979740143,
+      "eval_runtime": 25.4112,
+      "eval_samples_per_second": 8.067,
+      "eval_steps_per_second": 1.023,
+      "eval_wer": 0.3073237508555784,
+      "step": 300
+    },
+    {
+      "epoch": 6.85,
+      "eval_loss": 0.23940488696098328,
+      "eval_runtime": 25.3486,
+      "eval_samples_per_second": 8.087,
+      "eval_steps_per_second": 1.026,
+      "eval_wer": 0.2464065708418891,
+      "step": 350
+    },
+    {
+      "epoch": 7.83,
+      "eval_loss": 0.22206060588359833,
+      "eval_runtime": 25.3043,
+      "eval_samples_per_second": 8.101,
+      "eval_steps_per_second": 1.027,
+      "eval_wer": 0.22587268993839835,
+      "step": 400
+    },
+    {
+      "epoch": 8.82,
+      "eval_loss": 0.213778555393219,
+      "eval_runtime": 25.4337,
+      "eval_samples_per_second": 8.06,
+      "eval_steps_per_second": 1.022,
+      "eval_wer": 0.22039698836413416,
+      "step": 450
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 0.00012090225563909773,
+      "loss": 4.0907,
+      "step": 500
+    },
+    {
+      "epoch": 9.8,
+      "eval_loss": 0.2070361226797104,
+      "eval_runtime": 25.2632,
+      "eval_samples_per_second": 8.115,
+      "eval_steps_per_second": 1.029,
+      "eval_wer": 0.21765913757700206,
+      "step": 500
+    },
+    {
+      "epoch": 10.78,
+      "eval_loss": 0.2013179063796997,
+      "eval_runtime": 25.4168,
+      "eval_samples_per_second": 8.066,
+      "eval_steps_per_second": 1.023,
+      "eval_wer": 0.20123203285420946,
+      "step": 550
+    },
+    {
+      "epoch": 11.76,
+      "eval_loss": 0.2033400684595108,
+      "eval_runtime": 25.6076,
+      "eval_samples_per_second": 8.005,
+      "eval_steps_per_second": 1.015,
+      "eval_wer": 0.20465434633812457,
+      "step": 600
+    },
+    {
+      "epoch": 12.74,
+      "eval_loss": 0.20385989546775818,
+      "eval_runtime": 25.6153,
+      "eval_samples_per_second": 8.003,
+      "eval_steps_per_second": 1.015,
+      "eval_wer": 0.2026009582477755,
+      "step": 650
+    },
+    {
+      "epoch": 13.72,
+      "eval_loss": 0.2015053778886795,
+      "eval_runtime": 25.3511,
+      "eval_samples_per_second": 8.086,
+      "eval_steps_per_second": 1.026,
+      "eval_wer": 0.2053388090349076,
+      "step": 700
+    },
+    {
+      "epoch": 14.7,
+      "eval_loss": 0.19900637865066528,
+      "eval_runtime": 25.0763,
+      "eval_samples_per_second": 8.175,
+      "eval_steps_per_second": 1.037,
+      "eval_wer": 0.20191649555099248,
+      "step": 750
+    },
+    {
+      "epoch": 14.99,
+      "step": 765,
+      "total_flos": 5.601028718005862e+18,
+      "train_loss": 2.7169430053312014,
+      "train_runtime": 6457.8468,
+      "train_samples_per_second": 3.828,
+      "train_steps_per_second": 0.118
+    }
+  ],
+  "max_steps": 765,
+  "num_train_epochs": 15,
+  "total_flos": 5.601028718005862e+18,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:542d533cc085dedfa18d1cf1ce78830cdcb8ebb56784d0241bd4ddedf327c448
+size 2863

vocab.json

@@ -0,0 +1 @@
+{"\"": 1, ":": 2, ";": 3, "ᏸ": 4, "ᏹ": 5, "ᏺ": 6, "ᏻ": 7, "ᏼ": 8, "ꭰ": 9, "ꭱ": 10, "ꭲ": 11, "ꭳ": 12, "ꭴ": 13, "ꭵ": 14, "ꭶ": 15, "ꭷ": 16, "ꭸ": 17, "ꭹ": 18, "ꭺ": 19, "ꭻ": 20, "ꭼ": 21, "ꭽ": 22, "ꭾ": 23, "ꭿ": 24, "ꮀ": 25, "ꮁ": 26, "ꮂ": 27, "ꮃ": 28, "ꮄ": 29, "ꮅ": 30, "ꮆ": 31, "ꮇ": 32, "ꮈ": 33, "ꮉ": 34, "ꮊ": 35, "ꮋ": 36, "ꮌ": 37, "ꮎ": 38, "ꮏ": 39, "ꮑ": 40, "ꮒ": 41, "ꮓ": 42, "ꮔ": 43, "ꮕ": 44, "ꮖ": 45, "ꮗ": 46, "ꮘ": 47, "ꮙ": 48, "ꮚ": 49, "ꮛ": 50, "ꮜ": 51, "ꮝ": 52, "ꮞ": 53, "ꮟ": 54, "ꮠ": 55, "ꮡ": 56, "ꮢ": 57, "ꮣ": 58, "ꮤ": 59, "ꮥ": 60, "ꮦ": 61, "ꮧ": 62, "ꮨ": 63, "ꮩ": 64, "ꮪ": 65, "ꮫ": 66, "ꮭ": 67, "ꮮ": 68, "ꮯ": 69, "ꮰ": 70, "ꮱ": 71, "ꮲ": 72, "ꮳ": 73, "ꮴ": 74, "ꮵ": 75, "ꮶ": 76, "ꮷ": 77, "ꮸ": 78, "ꮹ": 79, "ꮺ": 80, "ꮻ": 81, "ꮼ": 82, "ꮽ": 83, "ꮾ": 84, "ꮿ": 85, "|": 0, "[UNK]": 86, "[PAD]": 87}