youfengy commited on
Commit
ab9ba71
·
verified ·
1 Parent(s): 4b027b1

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 sha256:806882bdfaedb5c778649e141cf26a43aff03651bdc44c2c34598fdd378a41d9
3
  size 166670441
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:abcbd255f9cea1d7f471bd4045f0ca2f0539b2c198385359cbaddaffbc1aa8a7
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:d93b9047daf8b9d59bece6835db9e3446d6f22fb1370c4971817a47b64aa7a66
3
  size 166832634
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f9ba58498554610131a5b1ef2ffd863eccce47f53aaa3564f0076cf36cf5544
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 sha256:0f83a81ca3f7080dc56cda4b2475e187b24bbbb88441d0ac49d5b64a236d960c
3
  size 42431444
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:029d562b5730c3d5eacd3b54ff6e3ccb0e915d4168db38297cfa575ba9aacb32
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:e267b6579d8cd84b51ffc01568986155b31b9b5bf9f897e9befaad6b250e806f
3
  size 42154677
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8485d3767d73d9359ba53b2b3ad8cbe154039f5a887d3667566fec0fd899a76f
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 sha256:9c68cbec12b5ea62b44380d7e04491127c779cab32ffe8b45e4f8344d6040307
3
  size 154076576
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60759e9f8b51648be9169c7627f42fde5306a8b61637601910d4b0dbc1147797
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:d4468c1111b535f970012a1837ec1bc0ced64a1bda982cb2c2d247d1dd19b32d
3
  size 38952139
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c95eba1f7d4b493702aabb091c65899e68818c5b713920032ba17129a8b51c6
3
  size 38952139
onnx/encoder_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:18909e4dc5ca2e559e40a62fa176849e9aa3a9e2a8cb34b3f9aaa27dc257dbd6
3
  size 141404308
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e8ad8b79fade96fc64fdff4ef91fc5fc745c2bcf1e9232178cec77e8912fb547
3
  size 141404308
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ac40dc8f57c4c9bac2ad485c84d210b2ced8dbc160cd2beb5df9babe6f21f78
3
  size 35640025
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e709a1d56783c36777f8dc46a2d9c3f70e0b03468eea02498c555a93d1b205a7
3
  size 35640025
quantize_config.json CHANGED
@@ -4,134 +4,134 @@
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
  }
 
4
  "per_model_config": {
5
  "decoder_model": {
6
  "op_types": [
7
+ "Pow",
 
8
  "Expand",
9
+ "Sub",
10
+ "Transpose",
11
  "Constant",
12
+ "Concat",
13
+ "Where",
14
+ "Log",
15
+ "Min",
16
+ "Shape",
17
  "Reshape",
18
+ "Sqrt",
19
  "Gather",
20
+ "Add",
 
21
  "Neg",
 
 
 
 
 
 
 
 
22
  "Tile",
23
+ "Div",
24
+ "Less",
 
 
25
  "Mul",
26
+ "Range",
27
+ "ReduceMean",
28
+ "Unsqueeze",
29
+ "ConstantOfShape",
30
+ "LessOrEqual",
31
+ "Cast",
32
+ "MatMul",
33
+ "Relu",
34
+ "Softmax"
35
  ],
36
  "weight_type": "QInt8"
37
  },
38
  "decoder_model_merged": {
39
  "op_types": [
40
+ "Pow",
 
41
  "Expand",
42
+ "Sub",
43
+ "Transpose",
44
  "Constant",
45
+ "Concat",
46
+ "Where",
47
+ "Log",
48
+ "Min",
49
+ "Shape",
50
  "Reshape",
51
+ "Sqrt",
52
  "Gather",
53
+ "Add",
 
54
  "Neg",
 
 
 
 
 
 
 
 
 
55
  "Tile",
56
+ "Div",
57
+ "Less",
58
+ "Mul",
59
+ "Range",
60
+ "If",
61
+ "ReduceMean",
62
+ "Unsqueeze",
63
  "ConstantOfShape",
64
+ "LessOrEqual",
65
  "Slice",
66
+ "Cast",
67
+ "MatMul",
68
+ "Relu",
69
+ "Softmax"
 
 
70
  ],
71
  "weight_type": "QInt8"
72
  },
73
  "decoder_with_past_model": {
74
  "op_types": [
75
+ "Pow",
 
76
  "Expand",
77
+ "Sub",
78
+ "Transpose",
79
  "Constant",
80
+ "Concat",
81
+ "Where",
82
+ "Log",
83
+ "Min",
84
+ "Shape",
85
  "Reshape",
86
+ "Sqrt",
87
  "Gather",
88
+ "Add",
 
89
  "Neg",
 
 
 
 
 
 
 
 
90
  "Tile",
91
+ "Div",
92
+ "Less",
93
+ "Mul",
94
+ "Range",
95
+ "ReduceMean",
96
+ "Unsqueeze",
97
  "ConstantOfShape",
98
+ "LessOrEqual",
99
  "Slice",
100
+ "Cast",
101
+ "MatMul",
102
+ "Relu",
103
+ "Softmax"
 
 
104
  ],
105
  "weight_type": "QInt8"
106
  },
107
  "encoder_model": {
108
  "op_types": [
109
+ "Pow",
110
+ "Sub",
111
+ "Transpose",
112
  "Constant",
113
+ "Concat",
114
+ "Where",
115
+ "Log",
116
+ "Min",
117
+ "Shape",
118
  "Reshape",
119
+ "Greater",
120
+ "Sqrt",
121
  "Gather",
122
+ "Abs",
123
+ "Add",
124
+ "Div",
125
+ "Less",
126
+ "Mul",
127
+ "Range",
128
  "ReduceMean",
129
+ "Unsqueeze",
130
+ "ConstantOfShape",
131
+ "Cast",
132
  "MatMul",
133
  "Relu",
134
+ "Softmax"
 
 
 
 
 
 
 
 
 
 
 
 
 
135
  ],
136
  "weight_type": "QInt8"
137
  }