Hugging Face's logo

onnx-community
/
lite-whisper-large-v3-turbo-acc-ONNX

Automatic Speech Recognition
Transformers.js
ONNX
lite-whisper
text2text-generation
Model card Files Community
Xenova HF staff commited on
Commit
73262f2
·
verified ·
1 Parent(s): 184424f

Upload folder using huggingface_hub

Browse files
Files changed (35) hide show
  1. config.json +258 -0
  2. generation_config.json +12 -0
  3. onnx/decoder_model.onnx +3 -0
  4. onnx/decoder_model_bnb4.onnx +3 -0
  5. onnx/decoder_model_fp16.onnx +3 -0
  6. onnx/decoder_model_int8.onnx +3 -0
  7. onnx/decoder_model_merged.onnx +3 -0
  8. onnx/decoder_model_merged_bnb4.onnx +3 -0
  9. onnx/decoder_model_merged_fp16.onnx +3 -0
  10. onnx/decoder_model_merged_int8.onnx +3 -0
  11. onnx/decoder_model_merged_q4.onnx +3 -0
  12. onnx/decoder_model_merged_q4f16.onnx +3 -0
  13. onnx/decoder_model_merged_quantized.onnx +3 -0
  14. onnx/decoder_model_merged_uint8.onnx +3 -0
  15. onnx/decoder_model_q4.onnx +3 -0
  16. onnx/decoder_model_q4f16.onnx +3 -0
  17. onnx/decoder_model_quantized.onnx +3 -0
  18. onnx/decoder_model_uint8.onnx +3 -0
  19. onnx/decoder_with_past_model.onnx +3 -0
  20. onnx/decoder_with_past_model_bnb4.onnx +3 -0
  21. onnx/decoder_with_past_model_fp16.onnx +3 -0
  22. onnx/decoder_with_past_model_int8.onnx +3 -0
  23. onnx/decoder_with_past_model_q4.onnx +3 -0
  24. onnx/decoder_with_past_model_q4f16.onnx +3 -0
  25. onnx/decoder_with_past_model_quantized.onnx +3 -0
  26. onnx/decoder_with_past_model_uint8.onnx +3 -0
  27. onnx/encoder_model.onnx +3 -0
  28. onnx/encoder_model_bnb4.onnx +3 -0
  29. onnx/encoder_model_fp16.onnx +3 -0
  30. onnx/encoder_model_int8.onnx +3 -0
  31. onnx/encoder_model_q4.onnx +3 -0
  32. onnx/encoder_model_q4f16.onnx +3 -0
  33. onnx/encoder_model_quantized.onnx +3 -0
  34. onnx/encoder_model_uint8.onnx +3 -0
  35. quantize_config.json +18 -0
config.json ADDED
@@ -0,0 +1,258 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": true,
3
+ "_name_or_path": "efficient-speech/lite-whisper-large-v3-turbo-acc",
4
+ "activation_dropout": 0.0,
5
+ "activation_function": "gelu",
6
+ "apply_spec_augment": false,
7
+ "architectures": [
8
+ "LiteWhisperForConditionalGeneration"
9
+ ],
10
+ "attention_dropout": 0.0,
11
+ "auto_map": {
12
+ "AutoConfig": "efficient-speech/lite-whisper-large-v3-turbo-acc--configuration_lite_whisper.LiteWhisperConfig",
13
+ "AutoModel": "efficient-speech/lite-whisper-large-v3-turbo-acc--modeling_lite_whisper.LiteWhisperForConditionalGeneration"
14
+ },
15
+ "begin_suppress_tokens": [
16
+ 220,
17
+ 50256
18
+ ],
19
+ "bos_token_id": 50257,
20
+ "classifier_proj_size": 256,
21
+ "d_model": 1280,
22
+ "decoder_attention_heads": 20,
23
+ "decoder_ffn_dim": 5120,
24
+ "decoder_layerdrop": 0.0,
25
+ "decoder_layers": 4,
26
+ "decoder_start_token_id": 50258,
27
+ "dropout": 0.0,
28
+ "encoder_attention_heads": 20,
29
+ "encoder_ffn_dim": 5120,
30
+ "encoder_layerdrop": 0.0,
31
+ "encoder_layers": 32,
32
+ "eos_token_id": 50257,
33
+ "init_std": 0.02,
34
+ "is_encoder_decoder": true,
35
+ "low_rank_config": [
36
+ {
37
+ "fc1": 352,
38
+ "fc2": 304,
39
+ "k_proj": 48,
40
+ "out_proj": 48,
41
+ "q_proj": 48,
42
+ "v_proj": 64
43
+ },
44
+ {
45
+ "fc1": 240,
46
+ "fc2": 256,
47
+ "k_proj": 80,
48
+ "out_proj": 192,
49
+ "q_proj": 96,
50
+ "v_proj": 192
51
+ },
52
+ {
53
+ "fc1": 192,
54
+ "fc2": 144,
55
+ "k_proj": 64,
56
+ "out_proj": 224,
57
+ "q_proj": 80,
58
+ "v_proj": 240
59
+ },
60
+ {
61
+ "fc1": 256,
62
+ "fc2": 128,
63
+ "k_proj": 80,
64
+ "out_proj": 288,
65
+ "q_proj": 80,
66
+ "v_proj": 336
67
+ },
68
+ {
69
+ "fc1": 272,
70
+ "fc2": 192,
71
+ "k_proj": 96,
72
+ "out_proj": 224,
73
+ "q_proj": 96,
74
+ "v_proj": 368
75
+ },
76
+ {
77
+ "fc1": 272,
78
+ "fc2": 304,
79
+ "k_proj": 80,
80
+ "out_proj": 256,
81
+ "q_proj": 80,
82
+ "v_proj": 352
83
+ },
84
+ {
85
+ "fc1": 352,
86
+ "fc2": 320,
87
+ "k_proj": 128,
88
+ "out_proj": 368,
89
+ "q_proj": 144,
90
+ "v_proj": 576
91
+ },
92
+ {
93
+ "fc1": 368,
94
+ "fc2": 560,
95
+ "k_proj": 128,
96
+ "out_proj": 384,
97
+ "q_proj": 160,
98
+ "v_proj": 512
99
+ },
100
+ {
101
+ "fc1": 464,
102
+ "fc2": 672,
103
+ "k_proj": 144,
104
+ "out_proj": 416,
105
+ "q_proj": 176,
106
+ "v_proj": 544
107
+ },
108
+ {
109
+ "fc1": 464,
110
+ "fc2": 688,
111
+ "k_proj": 160,
112
+ "out_proj": 448,
113
+ "q_proj": 192,
114
+ "v_proj": 608
115
+ },
116
+ {
117
+ "fc1": 528,
118
+ "fc2": 720,
119
+ "k_proj": 96,
120
+ "out_proj": 432,
121
+ "q_proj": 128,
122
+ "v_proj": 576
123
+ },
124
+ {
125
+ "fc1": 512,
126
+ "fc2": 720,
127
+ "k_proj": 128,
128
+ "out_proj": 432,
129
+ "q_proj": 160,
130
+ "v_proj": 608
131
+ },
132
+ {
133
+ "fc1": 544,
134
+ "fc2": 800,
135
+ "k_proj": 128,
136
+ "out_proj": 432,
137
+ "q_proj": 160,
138
+ "v_proj": 640
139
+ },
140
+ {
141
+ "fc1": 624,
142
+ "fc2": 928,
143
+ "k_proj": 112,
144
+ "out_proj": 496,
145
+ "q_proj": 128,
146
+ "v_proj": 624
147
+ },
148
+ {
149
+ "fc1": 704,
150
+ "fc2": 880,
151
+ "k_proj": 144,
152
+ "out_proj": 560,
153
+ "q_proj": 176
154
+ },
155
+ {
156
+ "fc1": 592,
157
+ "fc2": 992,
158
+ "k_proj": 80,
159
+ "out_proj": 560,
160
+ "q_proj": 96
161
+ },
162
+ {
163
+ "fc1": 640,
164
+ "fc2": 1024,
165
+ "k_proj": 96,
166
+ "out_proj": 432,
167
+ "q_proj": 112,
168
+ "v_proj": 576
169
+ },
170
+ {
171
+ "fc1": 736,
172
+ "k_proj": 160,
173
+ "out_proj": 544,
174
+ "q_proj": 192
175
+ },
176
+ {
177
+ "fc1": 800,
178
+ "k_proj": 128,
179
+ "out_proj": 480,
180
+ "q_proj": 160
181
+ },
182
+ {
183
+ "fc1": 848,
184
+ "k_proj": 144,
185
+ "out_proj": 512,
186
+ "q_proj": 160
187
+ },
188
+ {
189
+ "fc1": 832,
190
+ "fc2": 16,
191
+ "k_proj": 160,
192
+ "out_proj": 608,
193
+ "q_proj": 192
194
+ },
195
+ {
196
+ "fc1": 976,
197
+ "k_proj": 224,
198
+ "q_proj": 240
199
+ },
200
+ {
201
+ "fc1": 960,
202
+ "k_proj": 272,
203
+ "q_proj": 288
204
+ },
205
+ {
206
+ "k_proj": 272,
207
+ "q_proj": 288
208
+ },
209
+ {
210
+ "k_proj": 368,
211
+ "q_proj": 384
212
+ },
213
+ {
214
+ "k_proj": 336,
215
+ "q_proj": 384
216
+ },
217
+ {
218
+ "k_proj": 400,
219
+ "q_proj": 464
220
+ },
221
+ {
222
+ "k_proj": 400,
223
+ "q_proj": 448
224
+ },
225
+ {
226
+ "k_proj": 432,
227
+ "q_proj": 496
228
+ },
229
+ {
230
+ "k_proj": 560,
231
+ "q_proj": 608
232
+ },
233
+ {
234
+ "k_proj": 608
235
+ },
236
+ {
237
+ "k_proj": 624
238
+ }
239
+ ],
240
+ "mask_feature_length": 10,
241
+ "mask_feature_min_masks": 0,
242
+ "mask_feature_prob": 0.0,
243
+ "mask_time_length": 10,
244
+ "mask_time_min_masks": 2,
245
+ "mask_time_prob": 0.05,
246
+ "max_source_positions": 1500,
247
+ "max_target_positions": 448,
248
+ "median_filter_width": 7,
249
+ "model_type": "lite-whisper",
250
+ "num_hidden_layers": 32,
251
+ "num_mel_bins": 128,
252
+ "pad_token_id": 50257,
253
+ "scale_embedding": false,
254
+ "transformers_version": "4.48.3",
255
+ "use_cache": true,
256
+ "use_weighted_layer_sum": false,
257
+ "vocab_size": 51866
258
+ }
generation_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "begin_suppress_tokens": [
4
+ 220,
5
+ 50256
6
+ ],
7
+ "bos_token_id": 50257,
8
+ "decoder_start_token_id": 50258,
9
+ "eos_token_id": 50257,
10
+ "pad_token_id": 50257,
11
+ "transformers_version": "4.48.3"
12
+ }
onnx/decoder_model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ecad88e74cb556a0c54ac58c4d3a53cff23080393c0b4cef2a61679439421180
3
+ size 687782706
onnx/decoder_model_bnb4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2717c2969770220a5a70c14cdad56343d156e062f00de86bcbadb05fb7fda8e8
3
+ size 327340989
onnx/decoder_model_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7dd78de991e2b8ca0ff2112596932564120ba45aa1e27437995965c46f3242d2
3
+ size 343968493
onnx/decoder_model_int8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:062ccacb0cb1b3a1f5a55a53a674781a4c65f904d6db0f1ebbfe9e8c59132044
3
+ size 437927549
onnx/decoder_model_merged.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2868f3a8bc8cd051257a1e45e94aff32063d2ee213041b8a159c7f190d6e4b6
3
+ size 687963426
onnx/decoder_model_merged_bnb4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e293b0682cfa332dfea1e9a85cbb2e95d7e5ca68dae72371bbc67d74e792f00d
3
+ size 327526397
onnx/decoder_model_merged_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb72496cac1bac9ba55aa9f2fee0fbfe443c7e74d4e738579dbed564cc5fbd85
3
+ size 661764234
onnx/decoder_model_merged_int8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68b8ac2ef78b2cec9c1c76627ddba75a96914afa517e5c1485bce2de1564c959
3
+ size 438149217
onnx/decoder_model_merged_q4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d65f1ad128cf0755851678fa669095eb2b66986d3d385349b29dfc4e1ab0d86
3
+ size 334079385
onnx/decoder_model_merged_q4f16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31e23869d0af766adf4e1d9d3df34a69e7e10651df960ec6230a7fd7809105d9
3
+ size 379152361
onnx/decoder_model_merged_quantized.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68b8ac2ef78b2cec9c1c76627ddba75a96914afa517e5c1485bce2de1564c959
3
+ size 438149217
onnx/decoder_model_merged_uint8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd7c8c3a1ce421062907d3378a679587a43b5370d9520c92baa625f21db7bdb1
3
+ size 438149238
onnx/decoder_model_q4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c7c64061648294a38b30fe6fc18a99acf295fb7aaf5dbe46e633ece6c2a6810
3
+ size 333894265
onnx/decoder_model_q4f16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a82c06ac1d9dde3998725287547b32cb7c660ac035ea98b889a47ef812b21a5f
3
+ size 193241747
onnx/decoder_model_quantized.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:062ccacb0cb1b3a1f5a55a53a674781a4c65f904d6db0f1ebbfe9e8c59132044
3
+ size 437927549
onnx/decoder_model_uint8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56447dbceeedf66cf8f442b60c2d5c8e6d4615a01a03e0cb857ea8af03481f61
3
+ size 437927570
onnx/decoder_with_past_model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31fa5a003bec2df08311888978edaed5f5a8515365bbba5ca7a701b91a8d00ac
3
+ size 635300182
onnx/decoder_with_past_model_bnb4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50706d780823271ef08bc59cc4bc463e74b3090d267f72a87cb4d487606cc500
3
+ size 319913217
onnx/decoder_with_past_model_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fbdbdc5632c84c557451db9aac610528710faefd520cf9e93ee0efc4fd93a9e4
3
+ size 317714162
onnx/decoder_with_past_model_int8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24ce119c386434d9833cdbbc407623c8937088b93078044d17f306e2fd953d1a
3
+ size 424758658
onnx/decoder_with_past_model_q4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b84b8bf1a093da5e08565ad5e252df944943408645a2ed10ff798f1d5ea1a33
3
+ size 325647357
onnx/decoder_with_past_model_q4f16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:544fa56ce4a28e620abf3c72b23a5665c5afea2678b3e63c4acb7d3161e86c1b
3
+ size 185827736
onnx/decoder_with_past_model_quantized.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24ce119c386434d9833cdbbc407623c8937088b93078044d17f306e2fd953d1a
3
+ size 424758658
onnx/decoder_with_past_model_uint8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4da4ae7c54881f5c9c081f89effba1073a4e1d7231e029b0700ae7addabc4472
3
+ size 424758674
onnx/encoder_model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f561ac4c1c1d8fe2e996c41e959b73123437e6326cbbcdd605a82c15e7187cc
3
+ size 1690446051
onnx/encoder_model_bnb4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8c742deaf79a16df1efdf15c6914faa5a3cbc6603955d34b824a15e76543000
3
+ size 265186611
onnx/encoder_model_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0c6dfbc62c75c554c94a90d2ba99892a9e2176ec21ab28b03ec860ad3ab6a83
3
+ size 845441260
onnx/encoder_model_int8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32036448a6dcab9b2cfcba59aecd419787b25f50e98efd462f38cd6fe1efef7b
3
+ size 430681262
onnx/encoder_model_q4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:149fdffbfd5b02a4b5e7707e7033590bcde03ab3b1dff96f2b5f16906a1febc5
3
+ size 292315030
onnx/encoder_model_q4f16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0dab723632e81a4803a2d03e25429c11fc6e6b52603fbf6a49d4ce635173c557
3
+ size 250549960
onnx/encoder_model_quantized.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b0fa7f99d9413acf979d8c9cb1d1dfd5727ec4e057f45d92228444449ce162b
3
+ size 430681406
onnx/encoder_model_uint8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b0fa7f99d9413acf979d8c9cb1d1dfd5727ec4e057f45d92228444449ce162b
3
+ size 430681406
quantize_config.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "modes": [
3
+ "fp16",
4
+ "q8",
5
+ "int8",
6
+ "uint8",
7
+ "q4",
8
+ "q4f16",
9
+ "bnb4"
10
+ ],
11
+ "per_channel": false,
12
+ "reduce_range": false,
13
+ "block_size": null,
14
+ "is_symmetric": true,
15
+ "accuracy_level": null,
16
+ "quant_type": 1,
17
+ "op_block_list": null
18
+ }