Hugging Face's logo

huuquyet
/
PhoWhisper-tiny

Automatic Speech Recognition
Transformers.js
PyTorch
ONNX
Vietnamese
whisper
audio
hf-asr-leaderboard
Model card Files Community
huuquyet commited on
Commit
6d19be0
•
1 Parent(s): f72af40

feat: Update converted models with config

Browse files
Files changed (8) hide show
  1. config.json +0 -1
  2. merges.txt +0 -1
  3. onnx/decoder_model_merged.onnx +2 -2
  4. onnx/decoder_model_merged_quantized.onnx +2 -2
  5. preprocessor_config.json +0 -0
  6. quantize_config.json +68 -68
  7. tokenizer_config.json +33 -4
  8. vocab.json +0 -0
config.json CHANGED
@@ -146,7 +146,6 @@
146
  50361,
147
  50362
148
  ],
149
- "torch_dtype": "float32",
150
  "transformers_version": "4.33.2",
151
  "use_cache": true,
152
  "use_weighted_layer_sum": false,
 
146
  50361,
147
  50362
148
  ],
 
149
  "transformers_version": "4.33.2",
150
  "use_cache": true,
151
  "use_weighted_layer_sum": false,
merges.txt CHANGED
@@ -49998,4 +49998,3 @@ ctic amente
49998
  Ġmanif ests
49999
  ĠPak istani
50000
  Ġ( '
50001
- åľ º
 
49998
  Ġmanif ests
49999
  ĠPak istani
50000
  Ġ( '
 
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6c169b8c330d4345cdf607e23dbff2664bad8c767c7ead33343248a3df773bf5
3
- size 118533120
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13f7f875d3fcbc9883bcbaf031597c900ec45a42e3cf56907fa6cf8f4b4df978
3
+ size 118533111
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3a36371f1eac6675aec784a6cdb35f29154d3cf95bd8969f434c40269f11112d
3
- size 30667401
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:820865ac5e5162fcdffc44692dc1f50af6f29a1421391add0227db3497698404
3
+ size 30667374
preprocessor_config.json CHANGED
The diff for this file is too large to render. See raw diff
 
quantize_config.json CHANGED
@@ -4,110 +4,110 @@
4
  "per_model_config": {
5
  "encoder_model": {
6
  "op_types": [
7
- "Erf",
 
8
  "Add",
9
- "Mul",
10
- "Constant",
11
- "Sub",
12
- "Div",
13
- "MatMul",
14
  "Concat",
15
- "ReduceMean",
 
 
 
16
  "Transpose",
 
 
17
  "Sqrt",
18
- "Pow",
19
- "Conv",
20
- "Gather",
21
  "Unsqueeze",
22
- "Softmax",
23
- "Shape",
24
- "Reshape"
25
  ],
26
  "weight_type": "QUInt8"
27
  },
28
  "decoder_model": {
29
  "op_types": [
30
- "Erf",
31
- "Squeeze",
32
- "Expand",
33
- "Sub",
34
  "Concat",
35
- "Sqrt",
 
 
36
  "Unsqueeze",
37
- "Range",
38
- "Equal",
39
  "Constant",
40
- "ConstantOfShape",
41
- "ReduceMean",
42
- "Transpose",
43
- "Gather",
44
- "Less",
45
  "Shape",
 
 
 
46
  "Add",
47
- "Slice",
48
- "Div",
49
- "Where",
50
- "Reshape",
51
- "Cast",
52
  "Mul",
53
- "MatMul",
54
- "Pow",
55
- "Softmax"
 
 
 
56
  ],
57
  "weight_type": "QInt8"
58
  },
59
  "decoder_with_past_model": {
60
  "op_types": [
61
- "Erf",
62
- "Add",
63
- "Mul",
64
- "Constant",
65
- "Slice",
66
- "Sub",
67
- "Div",
68
- "MatMul",
69
- "Concat",
70
  "ReduceMean",
71
- "Transpose",
72
- "Sqrt",
73
  "Pow",
74
- "Gather",
75
- "Unsqueeze",
76
  "Softmax",
 
 
 
 
 
 
77
  "Shape",
78
- "Reshape"
 
 
 
 
79
  ],
80
  "weight_type": "QInt8"
81
  },
82
  "decoder_model_merged": {
83
  "op_types": [
84
- "Erf",
85
- "Squeeze",
86
- "Expand",
87
- "Sub",
88
  "Concat",
89
- "Sqrt",
90
- "If",
 
91
  "Unsqueeze",
92
- "Range",
93
- "Equal",
94
  "Constant",
95
- "ConstantOfShape",
96
- "ReduceMean",
97
- "Transpose",
98
- "Gather",
99
- "Less",
100
  "Shape",
 
 
 
101
  "Add",
102
- "Slice",
103
- "Div",
104
- "Where",
105
- "Reshape",
106
- "Cast",
107
  "Mul",
108
- "MatMul",
109
- "Pow",
110
- "Softmax"
 
 
 
111
  ],
112
  "weight_type": "QInt8"
113
  }
 
4
  "per_model_config": {
5
  "encoder_model": {
6
  "op_types": [
7
+ "Gather",
8
+ "ReduceMean",
9
  "Add",
10
+ "Pow",
 
 
 
 
11
  "Concat",
12
+ "Erf",
13
+ "Softmax",
14
+ "Div",
15
+ "Conv",
16
  "Transpose",
17
+ "Mul",
18
+ "Shape",
19
  "Sqrt",
20
+ "MatMul",
 
 
21
  "Unsqueeze",
22
+ "Reshape",
23
+ "Sub",
24
+ "Constant"
25
  ],
26
  "weight_type": "QUInt8"
27
  },
28
  "decoder_model": {
29
  "op_types": [
 
 
 
 
30
  "Concat",
31
+ "Div",
32
+ "Cast",
33
+ "Slice",
34
  "Unsqueeze",
35
+ "MatMul",
 
36
  "Constant",
37
+ "Where",
38
+ "Expand",
39
+ "Softmax",
 
 
40
  "Shape",
41
+ "Gather",
42
+ "Pow",
43
+ "ConstantOfShape",
44
  "Add",
45
+ "Squeeze",
46
+ "Equal",
47
+ "Range",
48
+ "Less",
 
49
  "Mul",
50
+ "Sub",
51
+ "ReduceMean",
52
+ "Erf",
53
+ "Transpose",
54
+ "Sqrt",
55
+ "Reshape"
56
  ],
57
  "weight_type": "QInt8"
58
  },
59
  "decoder_with_past_model": {
60
  "op_types": [
61
+ "Gather",
 
 
 
 
 
 
 
 
62
  "ReduceMean",
63
+ "Add",
 
64
  "Pow",
65
+ "Concat",
 
66
  "Softmax",
67
+ "Erf",
68
+ "MatMul",
69
+ "Div",
70
+ "Slice",
71
+ "Mul",
72
+ "Unsqueeze",
73
  "Shape",
74
+ "Sqrt",
75
+ "Transpose",
76
+ "Reshape",
77
+ "Sub",
78
+ "Constant"
79
  ],
80
  "weight_type": "QInt8"
81
  },
82
  "decoder_model_merged": {
83
  "op_types": [
 
 
 
 
84
  "Concat",
85
+ "Div",
86
+ "Cast",
87
+ "Slice",
88
  "Unsqueeze",
89
+ "MatMul",
 
90
  "Constant",
91
+ "Where",
92
+ "Expand",
93
+ "Softmax",
 
 
94
  "Shape",
95
+ "Gather",
96
+ "Pow",
97
+ "ConstantOfShape",
98
  "Add",
99
+ "If",
100
+ "Squeeze",
101
+ "Equal",
102
+ "Range",
103
+ "Less",
104
  "Mul",
105
+ "Sub",
106
+ "ReduceMean",
107
+ "Erf",
108
+ "Transpose",
109
+ "Sqrt",
110
+ "Reshape"
111
  ],
112
  "weight_type": "QInt8"
113
  }
tokenizer_config.json CHANGED
@@ -12976,14 +12976,43 @@
12976
  "<|nocaptions|>",
12977
  "<|notimestamps|>"
12978
  ],
12979
- "bos_token": "<|endoftext|>",
 
 
 
 
 
 
 
12980
  "clean_up_tokenization_spaces": true,
12981
- "eos_token": "<|endoftext|>",
 
 
 
 
 
 
 
12982
  "errors": "replace",
12983
  "model_max_length": 1024,
12984
- "pad_token": "<|endoftext|>",
 
 
 
 
 
 
 
12985
  "processor_class": "WhisperProcessor",
12986
  "return_attention_mask": false,
12987
  "tokenizer_class": "WhisperTokenizer",
12988
- "unk_token": "<|endoftext|>"
 
 
 
 
 
 
 
 
12989
  }
 
12976
  "<|nocaptions|>",
12977
  "<|notimestamps|>"
12978
  ],
12979
+ "bos_token": {
12980
+ "__type": "AddedToken",
12981
+ "content": "<|endoftext|>",
12982
+ "lstrip": false,
12983
+ "normalized": true,
12984
+ "rstrip": false,
12985
+ "single_word": false
12986
+ },
12987
  "clean_up_tokenization_spaces": true,
12988
+ "eos_token": {
12989
+ "__type": "AddedToken",
12990
+ "content": "<|endoftext|>",
12991
+ "lstrip": false,
12992
+ "normalized": true,
12993
+ "rstrip": false,
12994
+ "single_word": false
12995
+ },
12996
  "errors": "replace",
12997
  "model_max_length": 1024,
12998
+ "pad_token": {
12999
+ "__type": "AddedToken",
13000
+ "content": "<|endoftext|>",
13001
+ "lstrip": false,
13002
+ "normalized": true,
13003
+ "rstrip": false,
13004
+ "single_word": false
13005
+ },
13006
  "processor_class": "WhisperProcessor",
13007
  "return_attention_mask": false,
13008
  "tokenizer_class": "WhisperTokenizer",
13009
+ "trust_remote_code": false,
13010
+ "unk_token": {
13011
+ "__type": "AddedToken",
13012
+ "content": "<|endoftext|>",
13013
+ "lstrip": false,
13014
+ "normalized": true,
13015
+ "rstrip": false,
13016
+ "single_word": false
13017
+ }
13018
  }
vocab.json CHANGED
The diff for this file is too large to render. See raw diff