Wav2Vec2 for hi with LM !!!

alphabet.json

@@ -0,0 +1 @@
+{"labels": ["\u0930", "'", "i", "W", "\u2019", "w", "\u0906", "\u0909", "h", "\u095b", "\u091d", "\u0911", "N", "r", "\u093e", "\u0940", "\u0942", "d", "\u093c", "\u0907", "y", "\u0935", "u", "\u0948", "R", "\u0908", "\u092b", "\u092c", "\u0918", "E", "a", "\u0910", "\u0922", "\u0916", "M", "o", "J", "\u0949", "V", "Z", "\u0903", "c", "\u0939", "\u092f", "\u0905", "n", "\u095d", "\u0936", "\u0924", "l", "s", "F", "S", "\u093f", "\u0923", "B", "\u0932", "\u0902", "\u0938", "g", "L", "\u092e", "|", "G", "\u0901", "t", "\u0937", "\u091a", "x", "\u092a", "\u095c", "\u0941", "A", "X", "T", "v", "\u0920", "\u0947", "\u094d", " ", "\u0921", "p", "\u0943", "j", "\u091e", "\u0926", "m", "\u0927", "\u0945", "\u091f", "f", "\u0913", "\u091b", "b", "\u0958", "\u090a", "\u090b", "O", "C", "H", "\u091c", "\u094b", "e", "k", "P", "\u0915", "\u0917", "\u092d", "K", "\u090f", "\u094c", "\u0914", "\u0925", "\u0928", "\u2047", "", "<s>", "</s>"], "is_bpe": false}

config.json

@@ -0,0 +1,108 @@
+{
+  "_name_or_path": "LegolasTheElf/Wav2Vec2_xls_r_300m_hi_final",
+  "activation_dropout": 0.0,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.1,
+  "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": 0.1,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.0,
+  "gradient_checkpointing": false,
+  "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_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.075,
+  "model_type": "wav2vec2",
+  "num_adapter_layers": 3,
+  "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,
+  "output_hidden_size": 1024,
+  "pad_token_id": 115,
+  "proj_codevector_dim": 768,
+  "tdnn_dilation": [
+    1,
+    2,
+    3,
+    1,
+    1
+  ],
+  "tdnn_dim": [
+    512,
+    512,
+    512,
+    512,
+    1500
+  ],
+  "tdnn_kernel": [
+    5,
+    3,
+    3,
+    1,
+    1
+  ],
+  "torch_dtype": "float32",
+  "transformers_version": "4.16.2",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 118,
+  "xvector_output_dim": 512
+}

language_model/5gram_correct_new.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:666e5431e26a652cbff52183f4659508f3cf7b0ff6bc8a8b9ebefe2db5e9ee54
+size 149996785

language_model/attrs.json

@@ -0,0 +1 @@
+{"alpha": 0.5, "beta": 1.5, "unk_score_offset": -10.0, "score_boundary": true}

preprocessor_config.json

@@ -0,0 +1,10 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "processor_class": "Wav2Vec2ProcessorWithLM",
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85abd203b9d5666bd41435170093cb2eac516e0dae37fa291826a38f2c8fd491
+size 1262407473

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": " ", "processor_class": "Wav2Vec2ProcessorWithLM", "special_tokens_map_file": "/root/.cache/huggingface/transformers/df2d78003269fda0a80626d7299dd371cd535b2023ae7a60b0d00780a63ee4ef.a21d51735cf8667bcd610f057e88548d5d6a381401f6b4501a8bc6c1a9dc8498", "tokenizer_file": null, "name_or_path": "LegolasTheElf/Wav2Vec2_xls_r_300m_hi_final", "tokenizer_class": "Wav2Vec2CTCTokenizer"}

vocab.json

@@ -0,0 +1 @@
+{"र": 0, "'": 1, "i": 2, "W": 3, "’": 4, "w": 5, "आ": 6, "उ": 7, "h": 8, "ज़": 9, "झ": 10, "ऑ": 11, "N": 12, "r": 13, "ा": 14, "ी": 15, "ू": 16, "d": 17, "़": 18, "इ": 19, "y": 20, "व": 21, "u": 22, "ै": 23, "R": 24, "ई": 25, "फ": 26, "ब": 27, "घ": 28, "E": 29, "a": 30, "ऐ": 31, "ढ": 32, "ख": 33, "M": 34, "o": 35, "J": 36, "ॉ": 37, "V": 38, "Z": 39, "ः": 40, "c": 41, "ह": 42, "य": 43, "अ": 44, "n": 45, "ढ़": 46, "श": 47, "त": 48, "l": 49, "s": 50, "F": 51, "S": 52, "ि": 53, "ण": 54, "B": 55, "ल": 56, "ं": 57, "स": 58, "g": 59, "L": 60, "म": 61, "|": 62, "G": 63, "ँ": 64, "t": 65, "ष": 66, "च": 67, "x": 68, "प": 69, "ड़": 70, "ु": 71, "A": 72, "X": 73, "T": 74, "v": 75, "ठ": 76, "े": 77, "्": 78, " ": 79, "ड": 80, "p": 81, "ृ": 82, "j": 83, "ञ": 84, "द": 85, "m": 86, "ध": 87, "ॅ": 88, "ट": 89, "f": 90, "ओ": 91, "छ": 92, "b": 93, "क़": 94, "ऊ": 95, "ऋ": 96, "O": 97, "C": 98, "H": 99, "ज": 100, "ो": 101, "e": 102, "k": 103, "P": 104, "क": 105, "ग": 106, "भ": 107, "K": 108, "ए": 109, "ौ": 110, "औ": 111, "थ": 112, "न": 113, "[UNK]": 114, "[PAD]": 115, "<s>": 116, "</s>": 117}