Xenova HF staff commited on
Commit
ed3e4a4
1 Parent(s): 866fa59

Upload folder using huggingface_hub

Browse files
onnx/decoder_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a34147806885ea37601ae6d91a2598a93acc78db07a06e637f2f1e6ef9c90699
3
- size 819581461
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b2ceb285d8c3b54dc13baad904077f9fc0b3d926ff193e4345c9bca9c2f8dc0
3
+ size 613700175
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2c15a6f30dad41c3265f6786dbd850b6c563d9e6cca0197a1517822ccf2cb44b
3
- size 819827967
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32a8b5910c88ced64a7f8507dcc1fa785fa612fc27797105a2b7daf1843b8b0b
3
+ size 613946766
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a8b2aad1811a49576bb1fae43142bd91fc90204065a363191268ee84a7cd28
3
- size 206825235
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a584dce621f77e39008055b6ef4ae9d3698241758e568edf3e5b0ad5fd6ccb2
3
+ size 155102167
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b704da825ee605a2576a88093fed2442978b00309d3a576337ae2fe1ac843fcf
3
- size 206437988
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:034156c203b082ef11ac78e46cfffc96dd2f83f2c6c5a1d16401eef97291ae5e
3
+ size 154715574
onnx/decoder_with_past_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:145e78c5a35e7bd92f57f58b17b9538c202ccd2e4b233b8f4deb2d91e3454a32
3
- size 769158110
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:009480ed2abb7b8a20f127de89cf202b60f719ccf642813c332a5672f4d1506d
3
+ size 563276824
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8615325c7835f7734c6c6ed53264dc85f9954db1f176697cf033fc596fc7f52c
3
- size 193674668
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58a6832ec3a2e328322bae85925ad82b71622f0a535ccca65fa88bf082ee7c41
3
+ size 141952254
quantize_config.json CHANGED
@@ -4,120 +4,120 @@
4
  "per_model_config": {
5
  "decoder_model": {
6
  "op_types": [
7
- "Cast",
8
- "Transpose",
9
- "Slice",
10
- "Concat",
11
  "Unsqueeze",
 
 
12
  "Where",
13
- "Pow",
14
- "Expand",
15
- "MatMul",
16
- "Sub",
17
  "Div",
18
- "Equal",
19
- "Mul",
20
- "Gather",
21
  "Constant",
22
- "Softmax",
 
 
 
23
  "Shape",
24
- "Squeeze",
25
- "ConstantOfShape",
26
- "Range",
27
- "Less",
28
- "Add",
29
  "Sqrt",
 
30
  "Reshape",
31
- "ReduceMean",
32
- "Erf"
 
 
 
 
 
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
- "decoder_model_merged": {
37
  "op_types": [
38
- "Cast",
39
- "If",
40
- "Transpose",
41
- "Slice",
42
- "Concat",
43
  "Unsqueeze",
 
44
  "Where",
45
- "Pow",
46
- "Expand",
47
- "MatMul",
48
- "Sub",
49
  "Div",
50
- "Equal",
51
- "Mul",
52
- "Gather",
53
  "Constant",
54
- "Softmax",
 
 
55
  "Shape",
56
- "Squeeze",
57
- "ConstantOfShape",
58
- "Range",
59
- "Less",
60
- "Add",
61
  "Sqrt",
 
62
  "Reshape",
63
- "ReduceMean",
64
- "Erf"
 
 
 
 
 
65
  ],
66
  "weight_type": "QInt8"
67
  },
68
- "decoder_with_past_model": {
69
  "op_types": [
70
- "Cast",
71
- "Transpose",
72
- "Concat",
73
  "Unsqueeze",
 
 
74
  "Where",
75
- "Pow",
76
- "Expand",
77
- "MatMul",
78
- "Sub",
79
  "Div",
80
- "Equal",
81
- "Mul",
82
- "Gather",
83
  "Constant",
84
- "Softmax",
 
 
 
85
  "Shape",
86
- "ConstantOfShape",
87
- "Range",
88
- "Add",
89
  "Sqrt",
 
90
  "Reshape",
91
- "ReduceMean",
92
- "Erf"
 
 
 
 
 
 
93
  ],
94
  "weight_type": "QInt8"
95
  },
96
- "encoder_model": {
97
  "op_types": [
98
- "Cast",
99
- "Transpose",
100
- "Concat",
101
  "Unsqueeze",
 
102
  "Where",
103
- "Pow",
104
- "Expand",
105
- "MatMul",
106
- "Sub",
107
  "Div",
108
- "Equal",
109
- "Mul",
110
- "Gather",
111
  "Constant",
112
- "Softmax",
 
 
113
  "Shape",
114
- "ConstantOfShape",
115
- "Range",
116
- "Add",
117
  "Sqrt",
 
118
  "Reshape",
119
- "ReduceMean",
120
- "Erf"
 
 
 
 
 
121
  ],
122
  "weight_type": "QInt8"
123
  }
 
4
  "per_model_config": {
5
  "decoder_model": {
6
  "op_types": [
7
+ "Range",
8
+ "Less",
9
+ "Pow",
 
10
  "Unsqueeze",
11
+ "ConstantOfShape",
12
+ "Squeeze",
13
  "Where",
 
 
 
 
14
  "Div",
15
+ "ReduceMean",
 
 
16
  "Constant",
17
+ "Cast",
18
+ "Slice",
19
+ "Erf",
20
+ "MatMul",
21
  "Shape",
22
+ "Softmax",
 
 
 
 
23
  "Sqrt",
24
+ "Sub",
25
  "Reshape",
26
+ "Transpose",
27
+ "Expand",
28
+ "Mul",
29
+ "Gather",
30
+ "Equal",
31
+ "Concat",
32
+ "Add"
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
+ "encoder_model": {
37
  "op_types": [
38
+ "Range",
39
+ "Pow",
 
 
 
40
  "Unsqueeze",
41
+ "ConstantOfShape",
42
  "Where",
 
 
 
 
43
  "Div",
44
+ "ReduceMean",
 
 
45
  "Constant",
46
+ "Cast",
47
+ "Erf",
48
+ "MatMul",
49
  "Shape",
50
+ "Softmax",
 
 
 
 
51
  "Sqrt",
52
+ "Sub",
53
  "Reshape",
54
+ "Transpose",
55
+ "Expand",
56
+ "Gather",
57
+ "Mul",
58
+ "Equal",
59
+ "Concat",
60
+ "Add"
61
  ],
62
  "weight_type": "QInt8"
63
  },
64
+ "decoder_model_merged": {
65
  "op_types": [
66
+ "Range",
67
+ "Less",
68
+ "Pow",
69
  "Unsqueeze",
70
+ "ConstantOfShape",
71
+ "Squeeze",
72
  "Where",
 
 
 
 
73
  "Div",
74
+ "ReduceMean",
 
 
75
  "Constant",
76
+ "Cast",
77
+ "Slice",
78
+ "Erf",
79
+ "MatMul",
80
  "Shape",
81
+ "Softmax",
 
 
82
  "Sqrt",
83
+ "Sub",
84
  "Reshape",
85
+ "Transpose",
86
+ "If",
87
+ "Expand",
88
+ "Mul",
89
+ "Gather",
90
+ "Equal",
91
+ "Concat",
92
+ "Add"
93
  ],
94
  "weight_type": "QInt8"
95
  },
96
+ "decoder_with_past_model": {
97
  "op_types": [
98
+ "Range",
99
+ "Pow",
 
100
  "Unsqueeze",
101
+ "ConstantOfShape",
102
  "Where",
 
 
 
 
103
  "Div",
104
+ "ReduceMean",
 
 
105
  "Constant",
106
+ "Cast",
107
+ "Erf",
108
+ "MatMul",
109
  "Shape",
110
+ "Softmax",
 
 
111
  "Sqrt",
112
+ "Sub",
113
  "Reshape",
114
+ "Transpose",
115
+ "Expand",
116
+ "Mul",
117
+ "Gather",
118
+ "Equal",
119
+ "Concat",
120
+ "Add"
121
  ],
122
  "weight_type": "QInt8"
123
  }