jonatasgrosman/exp_w2v2t_th_unispeech-sat_s658

.gitattributes

@@ -25,3 +25,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+transcriptions_cv7_test.json filter=lfs diff=lfs merge=lfs -text
+transcriptions_cv7_validation.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,17 @@
+---
+language:
+- th
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- th
+datasets:
+- mozilla-foundation/common_voice_7_0
+---
+# exp_w2v2t_th_unispeech-sat_s658
+
+Fine-tuned [microsoft/unispeech-sat-large](https://huggingface.co/microsoft/unispeech-sat-large) for speech recognition using the train split of [Common Voice 7.0 (th)](https://huggingface.co/datasets/mozilla-foundation/common_voice_7_0).
+When using this model, make sure that your speech input is sampled at 16kHz.
+
+This model has been fine-tuned by the [HuggingSound](https://github.com/jonatasgrosman/huggingsound) tool.
+    

config.json

@@ -0,0 +1,104 @@
+{
+  "_name_or_path": "microsoft/unispeech-sat-large",
+  "activation_dropout": 0.05,
+  "apply_spec_augment": true,
+  "architectures": [
+    "UniSpeechSatForCTC"
+  ],
+  "attention_dropout": 0.05,
+  "bos_token_id": 1,
+  "classifier_proj_size": 256,
+  "codevector_dim": 768,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": false,
+  "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": "sum",
+  "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": 0.05,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.05,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.05,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.05,
+  "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": 0.05,
+  "model_type": "unispeech-sat",
+  "num_attention_heads": 16,
+  "num_clusters": 504,
+  "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": 0,
+  "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
+  ],
+  "tokenizer_class": "Wav2Vec2CTCTokenizer",
+  "torch_dtype": "float32",
+  "transformers_version": "4.16.0.dev0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 60,
+  "xvector_output_dim": 512
+}

preprocessor_config.json

@@ -0,0 +1,10 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "processor_class": "Wav2Vec2Processor",
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c136ad5796f36b794ffc826aca45abf772f39ee9a9bf841f2b1288cbfdb1e77c
+size 1262156927

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": "|", "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "/tmp/tmp9qr8bk5a", "tokenizer_class": "Wav2Vec2CTCTokenizer", "processor_class": "Wav2Vec2Processor"}

transcriptions_cv7_test.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff91010c6a8a7eca553cee08f29ca9c15e4e36a4c03d8f8d983fbc6b541cfc2d
+size 28853216

transcriptions_cv7_validation.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf008fba42f6644331795a57143bf7fdaf47af7c636e9d3dd3ec6222f90d20b7
+size 29339316

vocab.json

@@ -0,0 +1 @@
+{"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "'": 5, "-": 6, "ก": 7, "ข": 8, "ค": 9, "ฆ": 10, "ง": 11, "จ": 12, "ฉ": 13, "ช": 14, "ซ": 15, "ฌ": 16, "ญ": 17, "ฎ": 18, "ฏ": 19, "ฐ": 20, "ฑ": 21, "ฒ": 22, "ณ": 23, "ด": 24, "ต": 25, "ถ": 26, "ท": 27, "ธ": 28, "น": 29, "บ": 30, "ป": 31, "ผ": 32, "ฝ": 33, "พ": 34, "ฟ": 35, "ภ": 36, "ม": 37, "ย": 38, "ร": 39, "ฤ": 40, "ล": 41, "ว": 42, "ศ": 43, "ษ": 44, "ส": 45, "ห": 46, "ฬ": 47, "อ": 48, "ฮ": 49, "ะ": 50, "า": 51, "ำ": 52, "เ": 53, "แ": 54, "โ": 55, "ใ": 56, "ไ": 57, "ๅ": 58, "ๆ": 59}