DipsankarSinha commited on
Commit
0541e1d
·
verified ·
1 Parent(s): a461bce

Upload tokenizer

Browse files
Files changed (2) hide show
  1. README.md +6 -6
  2. vocab.json +228 -228
README.md CHANGED
@@ -1,18 +1,18 @@
1
  ---
2
- license: apache-2.0
3
  base_model: facebook/wav2vec2-xls-r-300m
4
- tags:
5
- - generated_from_trainer
6
  datasets:
7
  - common_voice_16_1
 
8
  metrics:
9
  - wer
 
 
10
  model-index:
11
  - name: wav2vec2-large-xls-r-300m-amharic-demo-colab
12
  results:
13
  - task:
14
- name: Automatic Speech Recognition
15
  type: automatic-speech-recognition
 
16
  dataset:
17
  name: common_voice_16_1
18
  type: common_voice_16_1
@@ -20,9 +20,9 @@ model-index:
20
  split: test
21
  args: am
22
  metrics:
23
- - name: Wer
24
- type: wer
25
  value: 0.8992661774516344
 
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
1
  ---
 
2
  base_model: facebook/wav2vec2-xls-r-300m
 
 
3
  datasets:
4
  - common_voice_16_1
5
+ license: apache-2.0
6
  metrics:
7
  - wer
8
+ tags:
9
+ - generated_from_trainer
10
  model-index:
11
  - name: wav2vec2-large-xls-r-300m-amharic-demo-colab
12
  results:
13
  - task:
 
14
  type: automatic-speech-recognition
15
+ name: Automatic Speech Recognition
16
  dataset:
17
  name: common_voice_16_1
18
  type: common_voice_16_1
 
20
  split: test
21
  args: am
22
  metrics:
23
+ - type: wer
 
24
  value: 0.8992661774516344
25
+ name: Wer
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
vocab.json CHANGED
@@ -1,232 +1,232 @@
1
  {
2
  "[PAD]": 229,
3
  "[UNK]": 228,
4
- "|": 85,
5
- "ሀ": 203,
6
- "ሁ": 39,
7
- "ሂ": 72,
8
- "ሃ": 174,
9
- "ሄ": 219,
10
- "ህ": 66,
11
- "ሆ": 184,
12
- "ለ": 183,
13
- "ሉ": 216,
14
- "ሊ": 98,
15
- "ላ": 166,
16
- "ሌ": 154,
17
- "ል": 187,
18
- "ሎ": 35,
19
- "ሏ": 114,
20
- "ሐ": 131,
21
- "ሑ": 49,
22
- "ሓ": 180,
23
- "ሔ": 132,
24
- "ሕ": 226,
25
- "መ": 198,
26
- "ሙ": 119,
27
- "ሚ": 196,
28
- "ማ": 59,
29
- "ሜ": 202,
30
- "ም": 57,
31
- "ሞ": 71,
32
- "ሟ": 0,
33
- "ሠ": 109,
34
- "ሡ": 122,
35
- "ሣ": 77,
36
- "ሥ": 116,
37
- "ሦ": 88,
38
- "ረ": 58,
39
- "ሩ": 70,
40
- "ሪ": 14,
41
- "ራ": 76,
42
- "ሬ": 204,
43
- "ር": 27,
44
- "ሮ": 171,
45
- "ሯ": 217,
46
- "ሰ": 212,
47
- "ሱ": 43,
48
- "ሲ": 163,
49
- "ሳ": 169,
50
- "ሴ": 56,
51
- "ስ": 182,
52
- "ሶ": 23,
53
- "ሷ": 103,
54
- "ሸ": 127,
55
- "ሹ": 106,
56
- "ሺ": 32,
57
- "ሻ": 101,
58
- "ሼ": 30,
59
- "ሽ": 197,
60
- "ሾ": 21,
61
- "ቀ": 210,
62
- "ቁ": 55,
63
- "ቂ": 65,
64
- "ቃ": 102,
65
- "ቄ": 188,
66
- "ቅ": 11,
67
- "ቆ": 74,
68
- "ቋ": 181,
69
- "በ": 111,
70
- "ቡ": 7,
71
- "ቢ": 207,
72
- "ባ": 139,
73
- "ቤ": 223,
74
- "ብ": 165,
75
- "ቦ": 100,
76
- "ቧ": 201,
77
- "ቨ": 159,
78
- "ቪ": 68,
79
- "ቫ": 199,
80
- "ቭ": 8,
81
- "ቮ": 41,
82
- "ተ": 82,
83
- "ቱ": 172,
84
- "ቲ": 34,
85
- "ታ": 177,
86
- "ቴ": 117,
87
- "ት": 225,
88
- "ቶ": 227,
89
- "ቷ": 37,
90
- "ቸ": 90,
91
- "ቹ": 200,
92
- "ቺ": 104,
93
- "ቻ": 141,
94
- "ቼ": 113,
95
- "ች": 209,
96
- "ቾ": 164,
97
- "ቿ": 222,
98
- "ኀ": 150,
99
- "ኃ": 52,
100
- "ኅ": 173,
101
- "ኋ": 220,
102
- "ነ": 193,
103
- "ኑ": 135,
104
- "ኒ": 143,
105
- "ና": 15,
106
- "ኔ": 44,
107
- "ን": 81,
108
- "ኖ": 148,
109
- "ኗ": 221,
110
- "ኘ": 208,
111
- "ኙ": 145,
112
- "ኛ": 170,
113
- "ኝ": 29,
114
- "ኞ": 5,
115
- "ኟ": 142,
116
- "አ": 3,
117
- "ኡ": 20,
118
- "ኢ": 110,
119
- "ኤ": 178,
120
- "እ": 33,
121
- "ኦ": 25,
122
- "ከ": 38,
123
- "ኩ": 153,
124
- "ኪ": 179,
125
- "ካ": 94,
126
- "ኬ": 136,
127
- "ክ": 138,
128
- "ኮ": 50,
129
- "ኳ": 147,
130
- "ኸ": 19,
131
- "ኽ": 31,
132
- "ወ": 97,
133
- "ዊ": 93,
134
- "ዋ": 121,
135
- "ዌ": 108,
136
- "ው": 214,
137
- "ዎ": 62,
138
- "ዐ": 186,
139
- "ዑ": 137,
140
- "ዒ": 2,
141
- "ዓ": 10,
142
- "ዕ": 124,
143
- "ዖ": 1,
144
- "ዘ": 73,
145
- "ዙ": 157,
146
- "ዚ": 206,
147
- "ዛ": 215,
148
- "ዜ": 130,
149
- "ዝ": 107,
150
- "ዞ": 60,
151
- "ዟ": 118,
152
- "ዢ": 140,
153
- "ዣ": 61,
154
- "ዤ": 91,
155
- "ዥ": 89,
156
- "ዦ": 112,
157
- "የ": 51,
158
- "ዩ": 167,
159
- "ያ": 128,
160
- "ዬ": 129,
161
- "ይ": 125,
162
- "ዮ": 47,
163
- "ደ": 195,
164
- "ዱ": 176,
165
- "ዲ": 156,
166
- "ዳ": 42,
167
- "ዴ": 84,
168
- "ድ": 69,
169
- "ዶ": 78,
170
- "ዷ": 146,
171
- "ጀ": 24,
172
- "ጁ": 158,
173
- "ጂ": 83,
174
- "ጃ": 155,
175
- "ጄ": 87,
176
- "ጅ": 22,
177
- "ጆ": 80,
178
- "ገ": 95,
179
- "ጉ": 189,
180
- "ጊ": 86,
181
- "ጋ": 211,
182
- "ጌ": 9,
183
- "ግ": 26,
184
- "ጎ": 99,
185
- "ጓ": 17,
186
- "ጠ": 126,
187
- "ጡ": 4,
188
- "ጢ": 133,
189
- "ጣ": 18,
190
- "ጤ": 205,
191
- "ጥ": 185,
192
- "ጦ": 96,
193
- "ጧ": 175,
194
- "ጨ": 67,
195
- "ጩ": 162,
196
- "ጪ": 79,
197
- "ጫ": 192,
198
- "ጬ": 40,
199
- "ጭ": 46,
200
- "ጮ": 218,
201
- "ጲ": 53,
202
- "ጴ": 161,
203
- "ጵ": 63,
204
- "ጶ": 168,
205
- "ጸ": 213,
206
- "ጹ": 54,
207
- "ጺ": 123,
208
- "ጻ": 160,
209
- "ጽ": 28,
210
- "ጾ": 115,
211
- "ጿ": 6,
212
- "ፀ": 12,
213
- "ፁ": 151,
214
- "ፃ": 48,
215
- "ፅ": 191,
216
- "ፈ": 64,
217
- "ፉ": 36,
218
- "ፊ": 92,
219
- "ፋ": 75,
220
- "ፌ": 190,
221
- "ፍ": 120,
222
- "ፎ": 224,
223
- "ፏ": 16,
224
- "ፑ": 45,
225
- "ፒ": 194,
226
- "ፓ": 134,
227
- "ፔ": 149,
228
- "ፕ": 13,
229
- "ፖ": 152,
230
- "“": 105,
231
- "”": 144
232
  }
 
1
  {
2
  "[PAD]": 229,
3
  "[UNK]": 228,
4
+ "|": 171,
5
+ "ሀ": 127,
6
+ "ሁ": 140,
7
+ "ሂ": 211,
8
+ "ሃ": 199,
9
+ "ሄ": 44,
10
+ "ህ": 86,
11
+ "ሆ": 180,
12
+ "ለ": 142,
13
+ "ሉ": 105,
14
+ "ሊ": 213,
15
+ "ላ": 32,
16
+ "ሌ": 111,
17
+ "ል": 172,
18
+ "ሎ": 67,
19
+ "ሏ": 207,
20
+ "ሐ": 158,
21
+ "ሑ": 63,
22
+ "ሓ": 159,
23
+ "ሔ": 60,
24
+ "ሕ": 118,
25
+ "መ": 203,
26
+ "ሙ": 54,
27
+ "ሚ": 16,
28
+ "ማ": 57,
29
+ "ሜ": 119,
30
+ "ም": 121,
31
+ "ሞ": 90,
32
+ "ሟ": 154,
33
+ "ሠ": 4,
34
+ "ሡ": 209,
35
+ "ሣ": 128,
36
+ "ሥ": 175,
37
+ "ሦ": 164,
38
+ "ረ": 191,
39
+ "ሩ": 101,
40
+ "ሪ": 227,
41
+ "ራ": 122,
42
+ "ሬ": 12,
43
+ "ር": 89,
44
+ "ሮ": 76,
45
+ "ሯ": 45,
46
+ "ሰ": 178,
47
+ "ሱ": 146,
48
+ "ሲ": 166,
49
+ "ሳ": 13,
50
+ "ሴ": 219,
51
+ "ስ": 80,
52
+ "ሶ": 107,
53
+ "ሷ": 109,
54
+ "ሸ": 28,
55
+ "ሹ": 169,
56
+ "ሺ": 198,
57
+ "ሻ": 167,
58
+ "ሼ": 78,
59
+ "ሽ": 36,
60
+ "ሾ": 37,
61
+ "ቀ": 5,
62
+ "ቁ": 216,
63
+ "ቂ": 145,
64
+ "ቃ": 70,
65
+ "ቄ": 147,
66
+ "ቅ": 38,
67
+ "ቆ": 94,
68
+ "ቋ": 185,
69
+ "በ": 88,
70
+ "ቡ": 71,
71
+ "ቢ": 168,
72
+ "ባ": 20,
73
+ "ቤ": 184,
74
+ "ብ": 33,
75
+ "ቦ": 17,
76
+ "ቧ": 6,
77
+ "ቨ": 218,
78
+ "ቪ": 222,
79
+ "ቫ": 27,
80
+ "ቭ": 196,
81
+ "ቮ": 160,
82
+ "ተ": 64,
83
+ "ቱ": 188,
84
+ "ቲ": 141,
85
+ "ታ": 62,
86
+ "ቴ": 96,
87
+ "ት": 72,
88
+ "ቶ": 52,
89
+ "ቷ": 77,
90
+ "ቸ": 40,
91
+ "ቹ": 42,
92
+ "ቺ": 197,
93
+ "ቻ": 176,
94
+ "ቼ": 14,
95
+ "ች": 92,
96
+ "ቾ": 34,
97
+ "ቿ": 215,
98
+ "ኀ": 144,
99
+ "ኃ": 59,
100
+ "ኅ": 148,
101
+ "ኋ": 217,
102
+ "ነ": 104,
103
+ "ኑ": 49,
104
+ "ኒ": 156,
105
+ "ና": 46,
106
+ "ኔ": 26,
107
+ "ን": 25,
108
+ "ኖ": 43,
109
+ "ኗ": 192,
110
+ "ኘ": 162,
111
+ "ኙ": 100,
112
+ "ኛ": 68,
113
+ "ኝ": 206,
114
+ "ኞ": 190,
115
+ "ኟ": 132,
116
+ "አ": 61,
117
+ "ኡ": 113,
118
+ "ኢ": 30,
119
+ "ኤ": 47,
120
+ "እ": 2,
121
+ "ኦ": 125,
122
+ "ከ": 117,
123
+ "ኩ": 69,
124
+ "ኪ": 31,
125
+ "ካ": 134,
126
+ "ኬ": 75,
127
+ "ክ": 194,
128
+ "ኮ": 173,
129
+ "ኳ": 83,
130
+ "ኸ": 138,
131
+ "ኽ": 201,
132
+ "ወ": 106,
133
+ "ዊ": 195,
134
+ "ዋ": 150,
135
+ "ዌ": 135,
136
+ "ው": 10,
137
+ "ዎ": 139,
138
+ "ዐ": 53,
139
+ "ዑ": 1,
140
+ "ዒ": 73,
141
+ "ዓ": 200,
142
+ "ዕ": 126,
143
+ "ዖ": 220,
144
+ "ዘ": 205,
145
+ "ዙ": 11,
146
+ "ዚ": 87,
147
+ "ዛ": 19,
148
+ "ዜ": 93,
149
+ "ዝ": 130,
150
+ "ዞ": 55,
151
+ "ዟ": 48,
152
+ "ዢ": 212,
153
+ "ዣ": 22,
154
+ "ዤ": 8,
155
+ "ዥ": 221,
156
+ "ዦ": 187,
157
+ "የ": 226,
158
+ "ዩ": 112,
159
+ "ያ": 202,
160
+ "ዬ": 182,
161
+ "ይ": 123,
162
+ "ዮ": 51,
163
+ "ደ": 223,
164
+ "ዱ": 174,
165
+ "ዲ": 208,
166
+ "ዳ": 3,
167
+ "ዴ": 177,
168
+ "ድ": 143,
169
+ "ዶ": 189,
170
+ "ዷ": 137,
171
+ "ጀ": 179,
172
+ "ጁ": 18,
173
+ "ጂ": 84,
174
+ "ጃ": 163,
175
+ "ጄ": 15,
176
+ "ጅ": 225,
177
+ "ጆ": 115,
178
+ "ገ": 35,
179
+ "ጉ": 81,
180
+ "ጊ": 65,
181
+ "ጋ": 157,
182
+ "ጌ": 74,
183
+ "ግ": 204,
184
+ "ጎ": 103,
185
+ "ጓ": 124,
186
+ "ጠ": 29,
187
+ "ጡ": 110,
188
+ "ጢ": 129,
189
+ "ጣ": 50,
190
+ "ጤ": 56,
191
+ "ጥ": 131,
192
+ "ጦ": 165,
193
+ "ጧ": 23,
194
+ "ጨ": 214,
195
+ "ጩ": 186,
196
+ "ጪ": 24,
197
+ "ጫ": 152,
198
+ "ጬ": 193,
199
+ "ጭ": 114,
200
+ "ጮ": 39,
201
+ "ጲ": 9,
202
+ "ጴ": 116,
203
+ "ጵ": 210,
204
+ "ጶ": 21,
205
+ "ጸ": 41,
206
+ "ጹ": 66,
207
+ "ጺ": 170,
208
+ "ጻ": 151,
209
+ "ጽ": 85,
210
+ "ጾ": 82,
211
+ "ጿ": 79,
212
+ "ፀ": 181,
213
+ "ፁ": 133,
214
+ "ፃ": 149,
215
+ "ፅ": 153,
216
+ "ፈ": 98,
217
+ "ፉ": 7,
218
+ "ፊ": 224,
219
+ "ፋ": 95,
220
+ "ፌ": 136,
221
+ "ፍ": 97,
222
+ "ፎ": 183,
223
+ "ፏ": 120,
224
+ "ፑ": 99,
225
+ "ፒ": 155,
226
+ "ፓ": 161,
227
+ "ፔ": 108,
228
+ "ፕ": 0,
229
+ "ፖ": 91,
230
+ "“": 58,
231
+ "”": 102
232
  }