jonatasgrosman commited on
Commit
8b0ad74
1 Parent(s): 3993c1a

jonatasgrosman/exp_w2v2t_th_unispeech-ml_s256

Browse files
.gitattributes CHANGED
@@ -25,3 +25,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
25
  *.zip filter=lfs diff=lfs merge=lfs -text
26
  *.zstandard filter=lfs diff=lfs merge=lfs -text
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
25
  *.zip filter=lfs diff=lfs merge=lfs -text
26
  *.zstandard filter=lfs diff=lfs merge=lfs -text
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
28
+ transcriptions_cv7_test.json filter=lfs diff=lfs merge=lfs -text
29
+ transcriptions_cv7_validation.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - th
4
+ license: apache-2.0
5
+ tags:
6
+ - automatic-speech-recognition
7
+ - th
8
+ datasets:
9
+ - mozilla-foundation/common_voice_7_0
10
+ ---
11
+ # exp_w2v2t_th_unispeech-ml_s256
12
+
13
+ Fine-tuned [microsoft/unispeech-large-multi-lingual-1500h-cv](https://huggingface.co/microsoft/unispeech-large-multi-lingual-1500h-cv) for speech recognition on Thai using the train split of [Common Voice 7.0](https://huggingface.co/datasets/mozilla-foundation/common_voice_7_0).
14
+ When using this model, make sure that your speech input is sampled at 16kHz.
15
+
16
+ This model has been fine-tuned by the [HuggingSound](https://github.com/jonatasgrosman/huggingsound) tool.
17
+
config.json ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/unispeech-large-multi-lingual-1500h-cv",
3
+ "activation_dropout": 0.05,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "UniSpeechForCTC"
7
+ ],
8
+ "attention_dropout": 0.05,
9
+ "bos_token_id": 1,
10
+ "classifier_proj_size": 256,
11
+ "codevector_dim": 768,
12
+ "contrastive_logits_temperature": 0.1,
13
+ "conv_bias": true,
14
+ "conv_dim": [
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512
22
+ ],
23
+ "conv_kernel": [
24
+ 10,
25
+ 3,
26
+ 3,
27
+ 3,
28
+ 3,
29
+ 2,
30
+ 2
31
+ ],
32
+ "conv_stride": [
33
+ 5,
34
+ 2,
35
+ 2,
36
+ 2,
37
+ 2,
38
+ 2,
39
+ 2
40
+ ],
41
+ "ctc_loss_reduction": "sum",
42
+ "ctc_zero_infinity": false,
43
+ "diversity_loss_weight": 0.1,
44
+ "do_stable_layer_norm": true,
45
+ "eos_token_id": 2,
46
+ "feat_extract_activation": "gelu",
47
+ "feat_extract_dropout": 0.0,
48
+ "feat_extract_norm": "layer",
49
+ "feat_proj_dropout": 0.05,
50
+ "feat_quantizer_dropout": 0.0,
51
+ "final_dropout": 0.05,
52
+ "hidden_act": "gelu",
53
+ "hidden_dropout": 0.05,
54
+ "hidden_size": 1024,
55
+ "initializer_range": 0.02,
56
+ "intermediate_size": 4096,
57
+ "layer_norm_eps": 1e-05,
58
+ "layerdrop": 0.05,
59
+ "mask_channel_length": 10,
60
+ "mask_channel_prob": 0.0,
61
+ "mask_feature_length": 10,
62
+ "mask_feature_min_masks": 0,
63
+ "mask_feature_prob": 0.0,
64
+ "mask_time_length": 10,
65
+ "mask_time_min_masks": 2,
66
+ "mask_time_min_space": 1,
67
+ "mask_time_prob": 0.05,
68
+ "model_type": "unispeech",
69
+ "num_attention_heads": 16,
70
+ "num_codevector_groups": 2,
71
+ "num_codevectors_per_group": 320,
72
+ "num_conv_pos_embedding_groups": 16,
73
+ "num_conv_pos_embeddings": 128,
74
+ "num_ctc_classes": 531,
75
+ "num_feat_extract_layers": 7,
76
+ "num_hidden_layers": 24,
77
+ "num_negatives": 100,
78
+ "pad_token_id": 0,
79
+ "proj_codevector_dim": 768,
80
+ "replace_prob": 0.5,
81
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
82
+ "torch_dtype": "float32",
83
+ "transformers_version": "4.16.0.dev0",
84
+ "use_weighted_layer_sum": false,
85
+ "vocab_size": 60
86
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d2276396a575a6c019d2a44b3d9c25194e2f350b0cd7af6cb8aa77a6656446f
3
+ size 1262170481
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
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/tmp9ib73knj", "tokenizer_class": "Wav2Vec2CTCTokenizer", "processor_class": "Wav2Vec2Processor"}
transcriptions_cv7_test.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2580e8d57146c37041c169cef958ffc44369bf793180e88d45ca55e9743662dd
3
+ size 28300488
transcriptions_cv7_validation.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b6a7d3268b1418db30738090be39af9a81e25ee9b3acc8d1c321f54a36f9b14
3
+ size 28663106
vocab.json ADDED
@@ -0,0 +1 @@
 
 
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}