youfengy commited on
Commit
4b027b1
1 Parent(s): 6082038

Upload 15 files

Browse files
onnx/decoder_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b17677a8e52be9f0a924f0e45f4c160aa5b90df9f8b5ea56ec37318f6b645044
3
  size 166670441
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:806882bdfaedb5c778649e141cf26a43aff03651bdc44c2c34598fdd378a41d9
3
  size 166670441
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8a5d16285704a1c060d4f88ed78c981647d27e8663b5fccf7d853f00b3a3bcaa
3
  size 166832634
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d93b9047daf8b9d59bece6835db9e3446d6f22fb1370c4971817a47b64aa7a66
3
  size 166832634
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:85822fdb4b29ab4b628f1028d4c488030b08eed68bbe37eadc6df9f72587ca19
3
  size 42431444
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f83a81ca3f7080dc56cda4b2475e187b24bbbb88441d0ac49d5b64a236d960c
3
  size 42431444
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ee7220db64f55f166559461cffc2141468af4615da0543bf659f65442e9fe17c
3
  size 42154677
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e267b6579d8cd84b51ffc01568986155b31b9b5bf9f897e9befaad6b250e806f
3
  size 42154677
onnx/decoder_with_past_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:56b85be5d484a1ea402fb68ee70cb9f10a8ebf31db582f4650c04c24d7d36aa5
3
  size 154076576
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c68cbec12b5ea62b44380d7e04491127c779cab32ffe8b45e4f8344d6040307
3
  size 154076576
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1412f385659a38913bb1c366f58dfd817f752283f911b64b7320d22c9a4973d5
3
  size 38952139
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4468c1111b535f970012a1837ec1bc0ced64a1bda982cb2c2d247d1dd19b32d
3
  size 38952139
onnx/encoder_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9bd48a64d314c8f74e541b676bbe9847ec4a24796a095b12e67fee6dda596cc5
3
  size 141404308
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18909e4dc5ca2e559e40a62fa176849e9aa3a9e2a8cb34b3f9aaa27dc257dbd6
3
  size 141404308
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:991e20004a6478ca0388c048dd6c440ea8f7c42208576519761146bdd3b20974
3
  size 35640025
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ac40dc8f57c4c9bac2ad485c84d210b2ced8dbc160cd2beb5df9babe6f21f78
3
  size 35640025
quantize_config.json CHANGED
@@ -4,134 +4,134 @@
4
  "per_model_config": {
5
  "decoder_model": {
6
  "op_types": [
 
7
  "Unsqueeze",
8
- "Sqrt",
9
- "ReduceMean",
10
- "Reshape",
11
  "Expand",
12
- "Shape",
13
- "Tile",
14
- "Transpose",
15
- "Gather",
16
- "Cast",
17
- "Range",
18
- "Where",
19
- "ConstantOfShape",
20
- "Concat",
21
  "Add",
22
  "Constant",
23
- "Sub",
24
- "Pow",
25
- "Less",
 
 
 
 
 
26
  "Relu",
27
- "Log",
28
  "Div",
29
- "Mul",
30
- "MatMul",
31
- "Neg",
32
  "Softmax",
33
- "LessOrEqual",
34
- "Min"
 
 
 
 
 
 
 
35
  ],
36
  "weight_type": "QInt8"
37
  },
38
  "decoder_model_merged": {
39
  "op_types": [
 
40
  "Unsqueeze",
41
- "Sqrt",
42
- "ReduceMean",
43
- "Reshape",
44
- "Slice",
45
  "Expand",
46
- "Shape",
47
- "If",
48
- "Tile",
49
- "Transpose",
50
- "Gather",
51
- "Cast",
52
- "Range",
53
- "Where",
54
- "ConstantOfShape",
55
- "Concat",
56
  "Add",
57
  "Constant",
58
- "Sub",
59
- "Pow",
60
- "Less",
 
 
 
 
 
61
  "Relu",
62
- "Log",
 
63
  "Div",
64
- "Mul",
65
- "MatMul",
66
- "Neg",
67
  "Softmax",
68
- "LessOrEqual",
69
- "Min"
 
 
 
 
 
 
 
 
70
  ],
71
  "weight_type": "QInt8"
72
  },
73
  "decoder_with_past_model": {
74
  "op_types": [
 
75
  "Unsqueeze",
76
- "Sqrt",
77
- "ReduceMean",
78
- "Reshape",
79
- "Slice",
80
  "Expand",
81
- "Shape",
82
- "Tile",
83
- "Transpose",
84
- "Gather",
85
- "Cast",
86
- "Range",
87
- "Where",
88
- "ConstantOfShape",
89
- "Concat",
90
  "Add",
91
  "Constant",
92
- "Sub",
93
- "Pow",
94
- "Less",
 
 
 
 
 
95
  "Relu",
96
- "Log",
97
  "Div",
98
- "Mul",
99
- "MatMul",
100
- "Neg",
101
  "Softmax",
102
- "LessOrEqual",
103
- "Min"
 
 
 
 
 
 
 
 
104
  ],
105
  "weight_type": "QInt8"
106
  },
107
  "encoder_model": {
108
  "op_types": [
109
- "Greater",
110
  "Unsqueeze",
111
- "Sqrt",
112
- "ReduceMean",
 
113
  "Reshape",
114
  "Abs",
115
- "Shape",
116
- "Transpose",
117
  "Gather",
118
- "Cast",
119
- "Range",
120
- "Where",
121
- "ConstantOfShape",
122
  "Concat",
123
- "Add",
124
- "Constant",
125
- "Sub",
126
- "Pow",
127
- "Less",
128
  "Relu",
129
- "Log",
130
  "Div",
131
- "Mul",
132
- "MatMul",
133
  "Softmax",
134
- "Min"
 
 
 
 
 
 
 
 
135
  ],
136
  "weight_type": "QInt8"
137
  }
 
4
  "per_model_config": {
5
  "decoder_model": {
6
  "op_types": [
7
+ "Range",
8
  "Unsqueeze",
 
 
 
9
  "Expand",
 
 
 
 
 
 
 
 
 
10
  "Add",
11
  "Constant",
12
+ "Cast",
13
+ "Reshape",
14
+ "LessOrEqual",
15
+ "Gather",
16
+ "ReduceMean",
17
+ "Concat",
18
+ "Neg",
19
+ "MatMul",
20
  "Relu",
21
+ "Sqrt",
22
  "Div",
23
+ "Shape",
24
+ "Pow",
 
25
  "Softmax",
26
+ "Min",
27
+ "Tile",
28
+ "ConstantOfShape",
29
+ "Transpose",
30
+ "Log",
31
+ "Where",
32
+ "Mul",
33
+ "Sub",
34
+ "Less"
35
  ],
36
  "weight_type": "QInt8"
37
  },
38
  "decoder_model_merged": {
39
  "op_types": [
40
+ "Range",
41
  "Unsqueeze",
 
 
 
 
42
  "Expand",
 
 
 
 
 
 
 
 
 
 
43
  "Add",
44
  "Constant",
45
+ "Cast",
46
+ "Reshape",
47
+ "LessOrEqual",
48
+ "Gather",
49
+ "ReduceMean",
50
+ "Concat",
51
+ "Neg",
52
+ "MatMul",
53
  "Relu",
54
+ "Sqrt",
55
+ "If",
56
  "Div",
57
+ "Shape",
58
+ "Pow",
 
59
  "Softmax",
60
+ "Min",
61
+ "Tile",
62
+ "ConstantOfShape",
63
+ "Slice",
64
+ "Transpose",
65
+ "Log",
66
+ "Where",
67
+ "Mul",
68
+ "Sub",
69
+ "Less"
70
  ],
71
  "weight_type": "QInt8"
72
  },
73
  "decoder_with_past_model": {
74
  "op_types": [
75
+ "Range",
76
  "Unsqueeze",
 
 
 
 
77
  "Expand",
 
 
 
 
 
 
 
 
 
78
  "Add",
79
  "Constant",
80
+ "Cast",
81
+ "Reshape",
82
+ "LessOrEqual",
83
+ "Gather",
84
+ "ReduceMean",
85
+ "Concat",
86
+ "Neg",
87
+ "MatMul",
88
  "Relu",
89
+ "Sqrt",
90
  "Div",
91
+ "Shape",
92
+ "Pow",
 
93
  "Softmax",
94
+ "Min",
95
+ "Tile",
96
+ "ConstantOfShape",
97
+ "Slice",
98
+ "Transpose",
99
+ "Log",
100
+ "Where",
101
+ "Mul",
102
+ "Sub",
103
+ "Less"
104
  ],
105
  "weight_type": "QInt8"
106
  },
107
  "encoder_model": {
108
  "op_types": [
109
+ "Range",
110
  "Unsqueeze",
111
+ "Add",
112
+ "Constant",
113
+ "Cast",
114
  "Reshape",
115
  "Abs",
 
 
116
  "Gather",
117
+ "ReduceMean",
 
 
 
118
  "Concat",
119
+ "MatMul",
 
 
 
 
120
  "Relu",
121
+ "Sqrt",
122
  "Div",
123
+ "Shape",
124
+ "Pow",
125
  "Softmax",
126
+ "Min",
127
+ "ConstantOfShape",
128
+ "Transpose",
129
+ "Greater",
130
+ "Log",
131
+ "Where",
132
+ "Mul",
133
+ "Sub",
134
+ "Less"
135
  ],
136
  "weight_type": "QInt8"
137
  }