Add Model Files

all_results.json

@@ -0,0 +1,23 @@
+{
+    "epoch": 30.0,
+    "eval_loss": 0.3495674431324005,
+    "eval_mem_cpu_alloc_delta": 443444278,
+    "eval_mem_cpu_peaked_delta": 34247898,
+    "eval_mem_gpu_alloc_delta": 0,
+    "eval_mem_gpu_peaked_delta": 2035233792,
+    "eval_runtime": 473.1711,
+    "eval_samples": 4641,
+    "eval_samples_per_second": 9.808,
+    "eval_wer": 0.28846481029887144,
+    "init_mem_cpu_alloc_delta": 10140165,
+    "init_mem_cpu_peaked_delta": 18306,
+    "init_mem_gpu_alloc_delta": 1261939712,
+    "init_mem_gpu_peaked_delta": 0,
+    "train_mem_cpu_alloc_delta": 1543641,
+    "train_mem_cpu_peaked_delta": 478977252,
+    "train_mem_gpu_alloc_delta": 3779933184,
+    "train_mem_gpu_peaked_delta": 7771228672,
+    "train_runtime": 43862.7197,
+    "train_samples": 11106,
+    "train_samples_per_second": 0.238
+}

config.json

@@ -0,0 +1,76 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
+  "activation_dropout": 0.1,
+  "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": 44,
+  "transformers_version": "4.5.0.dev0",
+  "vocab_size": 45
+}

eval_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 30.0,
+    "eval_loss": 0.3495674431324005,
+    "eval_mem_cpu_alloc_delta": 443444278,
+    "eval_mem_cpu_peaked_delta": 34247898,
+    "eval_mem_gpu_alloc_delta": 0,
+    "eval_mem_gpu_peaked_delta": 2035233792,
+    "eval_runtime": 473.1711,
+    "eval_samples": 4641,
+    "eval_samples_per_second": 9.808,
+    "eval_wer": 0.28846481029887144
+}

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:bfbf97987832b6f75ea8a8a917ace6e33a29e65413e85df5ea0bc683793da1b3
+size 1262118359

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": "|"}

train_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 30.0,
+    "init_mem_cpu_alloc_delta": 10140165,
+    "init_mem_cpu_peaked_delta": 18306,
+    "init_mem_gpu_alloc_delta": 1261939712,
+    "init_mem_gpu_peaked_delta": 0,
+    "train_mem_cpu_alloc_delta": 1543641,
+    "train_mem_cpu_peaked_delta": 478977252,
+    "train_mem_gpu_alloc_delta": 3779933184,
+    "train_mem_gpu_peaked_delta": 7771228672,
+    "train_runtime": 43862.7197,
+    "train_samples": 11106,
+    "train_samples_per_second": 0.238
+}

trainer_state.json

@@ -0,0 +1,303 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 30.0,
+  "global_step": 10440,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0003,
+      "loss": NaN,
+      "step": 500
+    },
+    {
+      "epoch": 1.44,
+      "eval_loss": 2.3965017795562744,
+      "eval_runtime": 464.2782,
+      "eval_samples_per_second": 9.996,
+      "eval_wer": 0.9993478137582941,
+      "step": 500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 0.00028490945674044263,
+      "loss": NaN,
+      "step": 1000
+    },
+    {
+      "epoch": 2.87,
+      "eval_loss": 0.40979766845703125,
+      "eval_runtime": 469.7691,
+      "eval_samples_per_second": 9.879,
+      "eval_wer": 0.4245448874269835,
+      "step": 1000
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 0.0002698189134808853,
+      "loss": NaN,
+      "step": 1500
+    },
+    {
+      "epoch": 4.31,
+      "eval_loss": 0.32825562357902527,
+      "eval_runtime": 475.331,
+      "eval_samples_per_second": 9.764,
+      "eval_wer": 0.37052685305960414,
+      "step": 1500
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 0.00025472837022132795,
+      "loss": NaN,
+      "step": 2000
+    },
+    {
+      "epoch": 5.75,
+      "eval_loss": 0.30649319291114807,
+      "eval_runtime": 474.8166,
+      "eval_samples_per_second": 9.774,
+      "eval_wer": 0.35002552033119716,
+      "step": 2000
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 0.0002396378269617706,
+      "loss": NaN,
+      "step": 2500
+    },
+    {
+      "epoch": 7.18,
+      "eval_loss": 0.3162720799446106,
+      "eval_runtime": 475.8373,
+      "eval_samples_per_second": 9.753,
+      "eval_wer": 0.3463392502693813,
+      "step": 2500
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 0.00022454728370221327,
+      "loss": NaN,
+      "step": 3000
+    },
+    {
+      "epoch": 8.62,
+      "eval_loss": 0.30734783411026,
+      "eval_runtime": 475.1293,
+      "eval_samples_per_second": 9.768,
+      "eval_wer": 0.3295808994498951,
+      "step": 3000
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 0.00020945674044265593,
+      "loss": NaN,
+      "step": 3500
+    },
+    {
+      "epoch": 10.06,
+      "eval_loss": 0.32298463582992554,
+      "eval_runtime": 478.2831,
+      "eval_samples_per_second": 9.703,
+      "eval_wer": 0.32586627346452673,
+      "step": 3500
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 0.0001943661971830986,
+      "loss": NaN,
+      "step": 4000
+    },
+    {
+      "epoch": 11.49,
+      "eval_loss": 0.3256703317165375,
+      "eval_runtime": 476.0024,
+      "eval_samples_per_second": 9.75,
+      "eval_wer": 0.3196279702829921,
+      "step": 4000
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 0.00017927565392354125,
+      "loss": NaN,
+      "step": 4500
+    },
+    {
+      "epoch": 12.93,
+      "eval_loss": 0.32864469289779663,
+      "eval_runtime": 473.6083,
+      "eval_samples_per_second": 9.799,
+      "eval_wer": 0.32019508875404074,
+      "step": 4500
+    },
+    {
+      "epoch": 14.37,
+      "learning_rate": 0.0001641851106639839,
+      "loss": NaN,
+      "step": 5000
+    },
+    {
+      "epoch": 14.37,
+      "eval_loss": 0.32181525230407715,
+      "eval_runtime": 474.4211,
+      "eval_samples_per_second": 9.782,
+      "eval_wer": 0.31662224238643455,
+      "step": 5000
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 0.00014909456740442654,
+      "loss": NaN,
+      "step": 5500
+    },
+    {
+      "epoch": 15.8,
+      "eval_loss": 0.33631598949432373,
+      "eval_runtime": 477.6284,
+      "eval_samples_per_second": 9.717,
+      "eval_wer": 0.3121136505415981,
+      "step": 5500
+    },
+    {
+      "epoch": 17.24,
+      "learning_rate": 0.0001340040241448692,
+      "loss": NaN,
+      "step": 6000
+    },
+    {
+      "epoch": 17.24,
+      "eval_loss": 0.3525121510028839,
+      "eval_runtime": 482.4872,
+      "eval_samples_per_second": 9.619,
+      "eval_wer": 0.30533658481256737,
+      "step": 6000
+    },
+    {
+      "epoch": 18.68,
+      "learning_rate": 0.00011891348088531185,
+      "loss": NaN,
+      "step": 6500
+    },
+    {
+      "epoch": 18.68,
+      "eval_loss": 0.34829291701316833,
+      "eval_runtime": 477.1506,
+      "eval_samples_per_second": 9.726,
+      "eval_wer": 0.3055634322009868,
+      "step": 6500
+    },
+    {
+      "epoch": 20.11,
+      "learning_rate": 0.00010382293762575451,
+      "loss": NaN,
+      "step": 7000
+    },
+    {
+      "epoch": 20.11,
+      "eval_loss": 0.35680249333381653,
+      "eval_runtime": 474.1059,
+      "eval_samples_per_second": 9.789,
+      "eval_wer": 0.30780355016162875,
+      "step": 7000
+    },
+    {
+      "epoch": 21.55,
+      "learning_rate": 8.873239436619717e-05,
+      "loss": NaN,
+      "step": 7500
+    },
+    {
+      "epoch": 21.55,
+      "eval_loss": 0.36808809638023376,
+      "eval_runtime": 477.9995,
+      "eval_samples_per_second": 9.709,
+      "eval_wer": 0.3064424658311121,
+      "step": 7500
+    },
+    {
+      "epoch": 22.99,
+      "learning_rate": 7.364185110663983e-05,
+      "loss": NaN,
+      "step": 8000
+    },
+    {
+      "epoch": 22.99,
+      "eval_loss": 0.36075320839881897,
+      "eval_runtime": 479.4594,
+      "eval_samples_per_second": 9.68,
+      "eval_wer": 0.29960868825497644,
+      "step": 8000
+    },
+    {
+      "epoch": 24.43,
+      "learning_rate": 5.855130784708249e-05,
+      "loss": NaN,
+      "step": 8500
+    },
+    {
+      "epoch": 24.43,
+      "eval_loss": 0.36022624373435974,
+      "eval_runtime": 482.5145,
+      "eval_samples_per_second": 9.618,
+      "eval_wer": 0.29745363806499175,
+      "step": 8500
+    },
+    {
+      "epoch": 25.86,
+      "learning_rate": 4.346076458752515e-05,
+      "loss": NaN,
+      "step": 9000
+    },
+    {
+      "epoch": 25.86,
+      "eval_loss": 0.35646718740463257,
+      "eval_runtime": 476.5808,
+      "eval_samples_per_second": 9.738,
+      "eval_wer": 0.2930301139908127,
+      "step": 9000
+    },
+    {
+      "epoch": 27.3,
+      "learning_rate": 2.8370221327967802e-05,
+      "loss": NaN,
+      "step": 9500
+    },
+    {
+      "epoch": 27.3,
+      "eval_loss": 0.3554363548755646,
+      "eval_runtime": 478.8286,
+      "eval_samples_per_second": 9.692,
+      "eval_wer": 0.29130040265411444,
+      "step": 9500
+    },
+    {
+      "epoch": 28.74,
+      "learning_rate": 1.3279678068410461e-05,
+      "loss": NaN,
+      "step": 10000
+    },
+    {
+      "epoch": 28.74,
+      "eval_loss": 0.3562660217285156,
+      "eval_runtime": 478.4327,
+      "eval_samples_per_second": 9.7,
+      "eval_wer": 0.29016616571201725,
+      "step": 10000
+    },
+    {
+      "epoch": 30.0,
+      "step": 10440,
+      "total_flos": 4.532712593948488e+19,
+      "train_runtime": 43862.7197,
+      "train_samples_per_second": 0.238
+    }
+  ],
+  "max_steps": 10440,
+  "num_train_epochs": 30,
+  "total_flos": 4.532712593948488e+19,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4770aa94d94300eefadfa2e8c000ddedbf2959ebf582123b6b38386ff4156145
+size 2351

vocab.json

@@ -0,0 +1 @@
+{"ó": 0, "q": 1, "á": 2, "ê": 3, "ô": 4, "c": 5, "”": 6, "ú": 7, "r": 8, "f": 9, "a": 10, "é": 11, "ü": 12, "l": 13, "ã": 14, "k": 15, "p": 16, "n": 17, "b": 18, "v": 19, "j": 20, "o": 21, "à": 22, "e": 23, "õ": 24, "t": 25, "y": 26, "ç": 27, "“": 28, "â": 29, "í": 30, "h": 31, "z": 32, "d": 33, "s": 34, "u": 35, "i": 36, "g": 37, "w": 38, "x": 39, "ñ": 41, "m": 42, "|": 40, "[UNK]": 43, "[PAD]": 44}