Kofi24 commited on
Commit
005d8b7
·
verified ·
1 Parent(s): d95da19

Initial commit: CTranslate2 INT8 quantized Whisper Tiny model

Browse files
Files changed (6) hide show
  1. README.md +60 -0
  2. config.json +223 -0
  3. model.bin +3 -0
  4. preprocessor_config.json +0 -0
  5. tokenizer.json +0 -0
  6. vocabulary.json +0 -0
README.md ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: mit
5
+ tags:
6
+ - whisper
7
+ - ctranslate2
8
+ - speech-recognition
9
+ - automatic-speech-recognition
10
+ library_name: ctranslate2
11
+ ---
12
+
13
+ # Whisper Tiny - CTranslate2 (INT8 Quantized)
14
+
15
+ This is a CTranslate2 converted version of OpenAI's Whisper Tiny model, quantized to INT8 for faster CPU inference.
16
+
17
+ ## Model Details
18
+
19
+ - **Base Model**: openai/whisper-tiny
20
+ - **Quantization**: INT8
21
+ - **Framework**: CTranslate2
22
+ - **Optimized for**: CPU inference
23
+
24
+ ## Usage
25
+
26
+ ### Install dependencies
27
+ ```bash
28
+ pip install faster-whisper
29
+ ```
30
+
31
+ ### Python code
32
+ ```python
33
+ from faster_whisper import WhisperModel
34
+
35
+ # Load model
36
+ model = WhisperModel("Kofi24/asr-whisper-tiny-ct2-int8", device="cpu", compute_type="int8")
37
+
38
+ # Transcribe
39
+ segments, info = model.transcribe("audio.mp3", language="en")
40
+
41
+ for segment in segments:
42
+ print(f"[{segment.start:.2f}s -> {segment.end:.2f}s] {segment.text}")
43
+ ```
44
+
45
+ ## Performance
46
+
47
+ - Faster inference than the original model
48
+ - Optimized for CPU with INT8 quantization
49
+ - Suitable for real-time transcription on CPU
50
+
51
+ ## Citation
52
+ ```bibtex
53
+ @misc{whisper-ctranslate2,
54
+ author = {Your Name},
55
+ title = {Whisper Tiny CTranslate2 INT8},
56
+ year = {2025},
57
+ publisher = {Hugging Face},
58
+ howpublished = {\url{https://huggingface.co/Kofi24/asr-whisper-tiny-ct2-int8}}
59
+ }
60
+ ```
config.json ADDED
@@ -0,0 +1,223 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 2,
5
+ 2
6
+ ],
7
+ [
8
+ 3,
9
+ 0
10
+ ],
11
+ [
12
+ 3,
13
+ 2
14
+ ],
15
+ [
16
+ 3,
17
+ 3
18
+ ],
19
+ [
20
+ 3,
21
+ 4
22
+ ],
23
+ [
24
+ 3,
25
+ 5
26
+ ]
27
+ ],
28
+ "lang_ids": [
29
+ 50259,
30
+ 50260,
31
+ 50261,
32
+ 50262,
33
+ 50263,
34
+ 50264,
35
+ 50265,
36
+ 50266,
37
+ 50267,
38
+ 50268,
39
+ 50269,
40
+ 50270,
41
+ 50271,
42
+ 50272,
43
+ 50273,
44
+ 50274,
45
+ 50275,
46
+ 50276,
47
+ 50277,
48
+ 50278,
49
+ 50279,
50
+ 50280,
51
+ 50281,
52
+ 50282,
53
+ 50283,
54
+ 50284,
55
+ 50285,
56
+ 50286,
57
+ 50287,
58
+ 50288,
59
+ 50289,
60
+ 50290,
61
+ 50291,
62
+ 50292,
63
+ 50293,
64
+ 50294,
65
+ 50295,
66
+ 50296,
67
+ 50297,
68
+ 50298,
69
+ 50299,
70
+ 50300,
71
+ 50301,
72
+ 50302,
73
+ 50303,
74
+ 50304,
75
+ 50305,
76
+ 50306,
77
+ 50307,
78
+ 50308,
79
+ 50309,
80
+ 50310,
81
+ 50311,
82
+ 50312,
83
+ 50313,
84
+ 50314,
85
+ 50315,
86
+ 50316,
87
+ 50317,
88
+ 50318,
89
+ 50319,
90
+ 50320,
91
+ 50321,
92
+ 50322,
93
+ 50323,
94
+ 50324,
95
+ 50325,
96
+ 50326,
97
+ 50327,
98
+ 50328,
99
+ 50329,
100
+ 50330,
101
+ 50331,
102
+ 50332,
103
+ 50333,
104
+ 50334,
105
+ 50335,
106
+ 50336,
107
+ 50337,
108
+ 50338,
109
+ 50339,
110
+ 50340,
111
+ 50341,
112
+ 50342,
113
+ 50343,
114
+ 50344,
115
+ 50345,
116
+ 50346,
117
+ 50347,
118
+ 50348,
119
+ 50349,
120
+ 50350,
121
+ 50351,
122
+ 50352,
123
+ 50353,
124
+ 50354,
125
+ 50355,
126
+ 50356,
127
+ 50357
128
+ ],
129
+ "suppress_ids": [
130
+ 1,
131
+ 2,
132
+ 7,
133
+ 8,
134
+ 9,
135
+ 10,
136
+ 14,
137
+ 25,
138
+ 26,
139
+ 27,
140
+ 28,
141
+ 29,
142
+ 31,
143
+ 58,
144
+ 59,
145
+ 60,
146
+ 61,
147
+ 62,
148
+ 63,
149
+ 90,
150
+ 91,
151
+ 92,
152
+ 93,
153
+ 359,
154
+ 503,
155
+ 522,
156
+ 542,
157
+ 873,
158
+ 893,
159
+ 902,
160
+ 918,
161
+ 922,
162
+ 931,
163
+ 1350,
164
+ 1853,
165
+ 1982,
166
+ 2460,
167
+ 2627,
168
+ 3246,
169
+ 3253,
170
+ 3268,
171
+ 3536,
172
+ 3846,
173
+ 3961,
174
+ 4183,
175
+ 4667,
176
+ 6585,
177
+ 6647,
178
+ 7273,
179
+ 9061,
180
+ 9383,
181
+ 10428,
182
+ 10929,
183
+ 11938,
184
+ 12033,
185
+ 12331,
186
+ 12562,
187
+ 13793,
188
+ 14157,
189
+ 14635,
190
+ 15265,
191
+ 15618,
192
+ 16553,
193
+ 16604,
194
+ 18362,
195
+ 18956,
196
+ 20075,
197
+ 21675,
198
+ 22520,
199
+ 26130,
200
+ 26161,
201
+ 26435,
202
+ 28279,
203
+ 29464,
204
+ 31650,
205
+ 32302,
206
+ 32470,
207
+ 36865,
208
+ 42863,
209
+ 47425,
210
+ 49870,
211
+ 50254,
212
+ 50258,
213
+ 50358,
214
+ 50359,
215
+ 50360,
216
+ 50361,
217
+ 50362
218
+ ],
219
+ "suppress_ids_begin": [
220
+ 220,
221
+ 50257
222
+ ]
223
+ }
model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4db7f7aa3b50951377de932d9ad80a59d5bcb8348fcdadf66ea0e976ef343f7a
3
+ size 40520007
preprocessor_config.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
vocabulary.json ADDED
The diff for this file is too large to render. See raw diff