initial_commit

config.json

@@ -0,0 +1,79 @@
+{
+  "_name_or_path": "./cv-tamil/checkpoint-1500",
+  "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,
+  "gradient_checkpointing": false,
+  "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_feature_length": 10,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_prob": 0.05,
+  "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": 38,
+  "proj_codevector_dim": 768,
+  "torch_dtype": "float32",
+  "transformers_version": "4.11.3",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 39
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6bb943e56e2df5046ceb88efdd942b3593e07337ffce379d29429ca2e905450
+size 2490378769

preprocessor_config.json

@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "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:d8e13478625d4c166ee6eec41a5c132ec9de11427d76d89f03349838e0846276
+size 1262083569

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:052df295d6910707b016bd0c631851bdbf6b8b0af7e4593fab5a737ef60ec8de
+size 14567

scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:814a41e021585f347b17d77b7bcddf012c401a989a74f43c0bed747c7fc33fc6
+size 559

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3318b9e9903d37e6b98bbf5760c23bcd11a61f77893e5ab75fac4a48bb3eddf1
+size 623

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": "|", "tokenizer_class": "Wav2Vec2CTCTokenizer"}

trainer_state.json

@@ -0,0 +1,133 @@
+{
+  "best_metric": 0.34306156635284424,
+  "best_model_checkpoint": "XLS-R-tamil-phoneme-v2/checkpoint-1500",
+  "epoch": 1.0706512042818912,
+  "global_step": 1500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.97e-05,
+      "loss": 7.2256,
+      "step": 100
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.94e-05,
+      "loss": 3.0456,
+      "step": 200
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 8.939999999999999e-05,
+      "loss": 2.6409,
+      "step": 300
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001191,
+      "loss": 2.4601,
+      "step": 400
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00014879999999999998,
+      "loss": 1.3437,
+      "step": 500
+    },
+    {
+      "epoch": 0.36,
+      "eval_cer": 0.402669688703466,
+      "eval_loss": 0.73838210105896,
+      "eval_runtime": 807.8476,
+      "eval_samples_per_second": 14.915,
+      "eval_steps_per_second": 1.865,
+      "step": 500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00017849999999999997,
+      "loss": 0.449,
+      "step": 600
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00020819999999999996,
+      "loss": 0.3495,
+      "step": 700
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0002382,
+      "loss": 0.3063,
+      "step": 800
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0002676,
+      "loss": 0.3021,
+      "step": 900
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00029759999999999997,
+      "loss": 0.2685,
+      "step": 1000
+    },
+    {
+      "epoch": 0.71,
+      "eval_cer": 0.2228818998716303,
+      "eval_loss": 0.3911130428314209,
+      "eval_runtime": 1044.7679,
+      "eval_samples_per_second": 11.533,
+      "eval_steps_per_second": 1.442,
+      "step": 1000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.00029980158159597406,
+      "loss": 0.2701,
+      "step": 1100
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.00029959022286125085,
+      "loss": 0.2527,
+      "step": 1200
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00029937886412652765,
+      "loss": 0.2472,
+      "step": 1300
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00029916319194823863,
+      "loss": 0.2317,
+      "step": 1400
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00029894751976994966,
+      "loss": 0.2197,
+      "step": 1500
+    },
+    {
+      "epoch": 1.07,
+      "eval_cer": 0.20153040757381258,
+      "eval_loss": 0.34306156635284424,
+      "eval_runtime": 1175.3837,
+      "eval_samples_per_second": 10.251,
+      "eval_steps_per_second": 1.282,
+      "step": 1500
+    }
+  ],
+  "max_steps": 140100,
+  "num_train_epochs": 100,
+  "total_flos": 9.531198923471569e+18,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9acdc5429d12bbbe88d35bb75f0f9bcaaeee33f5ba29005a6b771b144dda45e
+size 2863

vocab.json

@@ -0,0 +1 @@
+{"e": 1, "j": 2, "்": 3, "n̪": 4, "ɻ": 5, "h": 6, "n": 7, "oː": 8, "ʋ": 9, "ʈ": 10, "p": 11, "o": 12, "a": 13, "t͡ʃ": 14, "ŋ": 15, "uː": 16, "l": 17, "ɾ": 18, "iː": 19, "t̪": 20, "aː": 21, "eː": 22, "i": 23, "ʂ": 24, "m": 25, "u": 26, "d͡ʒ": 27, "r": 28, "ɳ": 29, "k": 30, "ɲ": 31, "ɭ": 32, "kʂ": 33, "aj": 34, "aʋ": 35, "s": 36, "|": 0, "[UNK]": 37, "[PAD]": 38}