jonatasgrosman commited on
Commit
e66a3e4
1 Parent(s): 35fdbb6

update model

Browse files
Files changed (5) hide show
  1. README.md +13 -6
  2. config.json +11 -4
  3. preprocessor_config.json +1 -0
  4. pytorch_model.bin +2 -2
  5. vocab.json +1 -1
README.md CHANGED
@@ -24,10 +24,10 @@ model-index:
24
  metrics:
25
  - name: Test WER
26
  type: wer
27
- value: 10.07
28
  - name: Test CER
29
  type: cer
30
- value: 3.04
31
  ---
32
 
33
  # Wav2Vec2-Large-XLSR-53-Spanish
@@ -49,7 +49,7 @@ from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
49
 
50
  LANG_ID = "es"
51
  MODEL_ID = "jonatasgrosman/wav2vec2-large-xlsr-53-spanish"
52
- SAMPLES = 5
53
 
54
  test_dataset = load_dataset("common_voice", LANG_ID, split=f"test[:{SAMPLES}]")
55
 
@@ -86,6 +86,11 @@ for i, predicted_sentence in enumerate(predicted_sentences):
86
  | PARA VISITAR CONTACTAR PRIMERO CON LA DIRECCIÓN. | PARA VISITAR CONTACTAR PRIMERO CON LA DIRECCIÓN |
87
  | TRES | TRES |
88
  | REALIZÓ LOS ESTUDIOS PRIMARIOS EN FRANCIA, PARA CONTINUAR LUEGO EN ESPAÑA. | REALIZÓ LOS ESTUDIOS PRIMARIOS EN FRANCIA PARA CONTINUAR LUEGO EN ESPAÑA |
 
 
 
 
 
89
 
90
  ## Evaluation
91
 
@@ -102,9 +107,11 @@ LANG_ID = "es"
102
  MODEL_ID = "jonatasgrosman/wav2vec2-large-xlsr-53-spanish"
103
  DEVICE = "cuda"
104
 
105
- CHARS_TO_IGNORE = [",", "?", "¿", ".", "!", "¡", ";", ":", '""', "%", '"', "�", "ʿ", "·", "჻", "~", "՞",
106
  "؟", "،", "।", "॥", "«", "»", "„", "“", "”", "「", "」", "‘", "’", "《", "》", "(", ")", "[", "]",
107
- "=", "`", "_", "+", "<", ">", "…", "–", "°", "´", "ʾ", "‹", "›", "©", "®", "—", "→", "。"]
 
 
108
 
109
  test_dataset = load_dataset("common_voice", LANG_ID, split="test")
110
 
@@ -156,7 +163,7 @@ In the table below I report the Word Error Rate (WER) and the Character Error Ra
156
 
157
  | Model | WER | CER |
158
  | ------------- | ------------- | ------------- |
159
- | jonatasgrosman/wav2vec2-large-xlsr-53-spanish | **10.07%** | **3.04%** |
160
  | pcuenq/wav2vec2-large-xlsr-53-es | 10.55% | 3.20% |
161
  | facebook/wav2vec2-large-xlsr-53-spanish | 16.99% | 5.40% |
162
  | mrm8488/wav2vec2-large-xlsr-53-spanish | 19.20% | 5.96% |
24
  metrics:
25
  - name: Test WER
26
  type: wer
27
+ value: 8.81
28
  - name: Test CER
29
  type: cer
30
+ value: 2.70
31
  ---
32
 
33
  # Wav2Vec2-Large-XLSR-53-Spanish
49
 
50
  LANG_ID = "es"
51
  MODEL_ID = "jonatasgrosman/wav2vec2-large-xlsr-53-spanish"
52
+ SAMPLES = 10
53
 
54
  test_dataset = load_dataset("common_voice", LANG_ID, split=f"test[:{SAMPLES}]")
55
 
86
  | PARA VISITAR CONTACTAR PRIMERO CON LA DIRECCIÓN. | PARA VISITAR CONTACTAR PRIMERO CON LA DIRECCIÓN |
87
  | TRES | TRES |
88
  | REALIZÓ LOS ESTUDIOS PRIMARIOS EN FRANCIA, PARA CONTINUAR LUEGO EN ESPAÑA. | REALIZÓ LOS ESTUDIOS PRIMARIOS EN FRANCIA PARA CONTINUAR LUEGO EN ESPAÑA |
89
+ | EN LOS AÑOS QUE SIGUIERON, ESTE TRABAJO ESPARTA PRODUJO DOCENAS DE BUENOS JUGADORES. | EN LOS AÑOS QUE SIGUIERON ESTE TRABAJO ESPARTA PRODUJO DOCENA DE BUENOS JUGADORES |
90
+ | SE ESTÁ TRATANDO DE RECUPERAR SU CULTIVO EN LAS ISLAS CANARIAS. | SE ESTÓ TRATANDO DE RECUPERAR SU CULTIVO EN LAS ISLAS CANARIAS |
91
+ | SÍ | SÍ |
92
+ | "FUE ""SACADA"" DE LA SERIE EN EL EPISODIO ""LEAD"", EN QUE ALEXANDRA CABOT REGRESÓ." | FUE SACADA DE LA SERIE EN EL EPISODIO LEED EN QUE ALEXANDRA KAOT REGRESÓ |
93
+ | SE UBICAN ESPECÍFICAMENTE EN EL VALLE DE MOKA, EN LA PROVINCIA DE BIOKO SUR. | SE UBICAN ESPECÍFICAMENTE EN EL VALLE DE MOCA EN LA PROVINCIA DE PÍOCOSUR |
94
 
95
  ## Evaluation
96
 
107
  MODEL_ID = "jonatasgrosman/wav2vec2-large-xlsr-53-spanish"
108
  DEVICE = "cuda"
109
 
110
+ CHARS_TO_IGNORE = [",", "?", "¿", ".", "!", "¡", ";", ";", ":", '""', "%", '"', "�", "ʿ", "·", "჻", "~", "՞",
111
  "؟", "،", "।", "॥", "«", "»", "„", "“", "”", "「", "」", "‘", "’", "《", "》", "(", ")", "[", "]",
112
+ "{", "}", "=", "`", "_", "+", "<", ">", "…", "–", "°", "´", "ʾ", "‹", "›", "©", "®", "—", "→", "。",
113
+ "、", "﹂", "﹁", "‧", "~", "﹏", ",", "{", "}", "(", ")", "[", "]", "【", "】", "‥", "〽",
114
+ "『", "』", "〝", "〟", "⟨", "⟩", "〜", ":", "!", "?", "♪", "؛", "/", "\\", "º", "−", "^", "ʻ", "ˆ"]
115
 
116
  test_dataset = load_dataset("common_voice", LANG_ID, split="test")
117
 
163
 
164
  | Model | WER | CER |
165
  | ------------- | ------------- | ------------- |
166
+ | jonatasgrosman/wav2vec2-large-xlsr-53-spanish | **8.81%** | **2.70%** |
167
  | pcuenq/wav2vec2-large-xlsr-53-es | 10.55% | 3.20% |
168
  | facebook/wav2vec2-large-xlsr-53-spanish | 16.99% | 5.40% |
169
  | mrm8488/wav2vec2-large-xlsr-53-spanish | 19.20% | 5.96% |
config.json CHANGED
@@ -43,20 +43,27 @@
43
  "feat_extract_dropout": 0.0,
44
  "feat_extract_norm": "layer",
45
  "feat_proj_dropout": 0.05,
46
- "final_dropout": 0.1,
47
  "gradient_checkpointing": true,
48
  "hidden_act": "gelu",
49
  "hidden_dropout": 0.05,
50
- "hidden_dropout_prob": 0.1,
51
  "hidden_size": 1024,
52
  "initializer_range": 0.02,
53
  "intermediate_size": 4096,
54
  "layer_norm_eps": 1e-05,
55
  "layerdrop": 0.05,
 
 
 
 
 
56
  "mask_feature_length": 10,
57
  "mask_feature_prob": 0.0,
58
  "mask_time_length": 10,
 
 
59
  "mask_time_prob": 0.05,
 
60
  "model_type": "wav2vec2",
61
  "num_attention_heads": 16,
62
  "num_conv_pos_embedding_groups": 16,
@@ -64,6 +71,6 @@
64
  "num_feat_extract_layers": 7,
65
  "num_hidden_layers": 24,
66
  "pad_token_id": 0,
67
- "transformers_version": "4.5.0.dev0",
68
- "vocab_size": 40
69
  }
43
  "feat_extract_dropout": 0.0,
44
  "feat_extract_norm": "layer",
45
  "feat_proj_dropout": 0.05,
46
+ "final_dropout": 0.0,
47
  "gradient_checkpointing": true,
48
  "hidden_act": "gelu",
49
  "hidden_dropout": 0.05,
 
50
  "hidden_size": 1024,
51
  "initializer_range": 0.02,
52
  "intermediate_size": 4096,
53
  "layer_norm_eps": 1e-05,
54
  "layerdrop": 0.05,
55
+ "mask_channel_length": 10,
56
+ "mask_channel_min_space": 1,
57
+ "mask_channel_other": 0.0,
58
+ "mask_channel_prob": 0.0,
59
+ "mask_channel_selection": "static",
60
  "mask_feature_length": 10,
61
  "mask_feature_prob": 0.0,
62
  "mask_time_length": 10,
63
+ "mask_time_min_space": 1,
64
+ "mask_time_other": 0.0,
65
  "mask_time_prob": 0.05,
66
+ "mask_time_selection": "static",
67
  "model_type": "wav2vec2",
68
  "num_attention_heads": 16,
69
  "num_conv_pos_embedding_groups": 16,
71
  "num_feat_extract_layers": 7,
72
  "num_hidden_layers": 24,
73
  "pad_token_id": 0,
74
+ "transformers_version": "4.7.0.dev0",
75
+ "vocab_size": 41
76
  }
preprocessor_config.json CHANGED
@@ -1,5 +1,6 @@
1
  {
2
  "do_normalize": true,
 
3
  "feature_size": 1,
4
  "padding_side": "right",
5
  "padding_value": 0.0,
1
  {
2
  "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
  "feature_size": 1,
5
  "padding_side": "right",
6
  "padding_value": 0.0,
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:60c878351b0f005d0575b84427412dc7500ab0090e5513dd1e3b4343cb87aa4e
3
- size 1262097815
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41c110e55d2eac8c79486ad87dbe8f9527ed034fe087a6adf03c891eeba914c1
3
+ size 1262101911
vocab.json CHANGED
@@ -1 +1 @@
1
- {"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "E": 5, "A": 6, "O": 7, "S": 8, "N": 9, "R": 10, "L": 11, "I": 12, "D": 13, "U": 14, "T": 15, "C": 16, "M": 17, "P": 18, "B": 19, "Q": 20, "Y": 21, "H": 22, "G": 23, "V": 24, "Í": 25, "Á": 26, "F": 27, "Ó": 28, "J": 29, "É": 30, "Z": 31, "Ñ": 32, "X": 33, "Ú": 34, "'": 35, "K": 36, "W": 37, "Ü": 38, "-": 39}
1
+ {"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "'": 5, "-": 6, "A": 7, "B": 8, "C": 9, "D": 10, "E": 11, "F": 12, "G": 13, "H": 14, "I": 15, "J": 16, "K": 17, "L": 18, "M": 19, "N": 20, "O": 21, "P": 22, "Q": 23, "R": 24, "S": 25, "T": 26, "U": 27, "V": 28, "W": 29, "X": 30, "Y": 31, "Z": 32, "Á": 33, "É": 34, "Í": 35, "Ñ": 36, "Ó": 37, "Ö": 38, "Ú": 39, "Ü": 40}