Xenova HF staff commited on
Commit
3442592
1 Parent(s): 62ee5f0

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:d094736e8b3b6e67e9c3b8a30259a053f06e7bd75d5efd4ea406b6dc441f8690
3
- size 314523257
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08ca2300844d32d52d5743bf6ec9a2c0953a11f90a031f82b98c4c373813a529
3
+ size 208304195
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:eb8ca7ec5a9bcdc49011bc15946cf334a9cf908a5b0fff6d3165a0b178670913
3
- size 314779197
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5df8d95539f2f87ec81663f676aab11772b766bced6b4ba21735255a6f317c4
3
+ size 208560983
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4a6134ef1ab4b1d7fa2f0aadfd15478537db3fe86793c251abf65630591767f6
3
- size 80262105
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e2bd3a0083d3e31e9c3e29d089d1d44eb365a782781389f82fd5ed133f7589a
3
+ size 53707540
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ff30d094886804d49108f957ac895bef102352e310c960a58e0f8856f3dd5670
3
- size 79864827
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d83958fa9b5279484cdc3b5923c903450663683d60980f29fdb1fe4b64cbdd9d
3
+ size 53309878
onnx/decoder_with_past_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4adecec0758b3919d07f8d1127d3c34eb01e3f1ece2e874f60a3227ddd238dca
3
- size 301889972
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b90cf3b333790a022840a80a38947d5b3b262c9df9a0424af4aac2e617c984b3
3
+ size 195670910
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d1b67629c24bba0294ec5f054933dc4ed17deb2861f6b8973ffca2ec98af5c46
3
- size 76643187
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e46514e9301c3721972b6ddc044d2c37749374dda74fd0c9dc27b75c29d8e64
3
+ size 50088238
quantize_config.json CHANGED
@@ -2,112 +2,112 @@
2
  "per_channel": false,
3
  "reduce_range": false,
4
  "per_model_config": {
5
- "decoder_model_merged": {
6
  "op_types": [
 
7
  "Gather",
8
- "Shape",
9
- "Cast",
 
 
 
 
 
 
10
  "Reshape",
 
 
11
  "Sqrt",
12
- "Constant",
13
- "Unsqueeze",
14
- "Squeeze",
15
  "Pow",
16
  "Concat",
17
- "MatMul",
18
- "Erf",
19
- "Range",
20
  "Softmax",
 
21
  "Sub",
22
- "Less",
23
  "Div",
24
- "ReduceMean",
25
- "If",
26
- "Slice",
27
- "Where",
28
- "Add",
29
- "Mul",
30
- "Expand",
31
- "ConstantOfShape",
32
- "Transpose",
33
- "Equal"
34
  ],
35
  "weight_type": "QInt8"
36
  },
37
- "decoder_model": {
38
  "op_types": [
 
39
  "Gather",
40
- "Shape",
41
- "Cast",
 
 
 
 
 
 
42
  "Reshape",
 
 
43
  "Sqrt",
44
- "Constant",
45
- "Unsqueeze",
46
- "Squeeze",
47
  "Pow",
48
  "Concat",
49
- "MatMul",
50
- "Erf",
51
- "Range",
52
  "Softmax",
 
53
  "Sub",
54
- "Less",
55
  "Div",
56
- "ReduceMean",
57
- "Slice",
58
- "Where",
59
- "Add",
60
- "Mul",
61
- "Expand",
62
- "ConstantOfShape",
63
- "Transpose",
64
- "Equal"
65
  ],
66
  "weight_type": "QInt8"
67
  },
68
  "decoder_with_past_model": {
69
  "op_types": [
70
- "Unsqueeze",
71
- "Gather",
72
- "Shape",
73
- "Sub",
74
- "Div",
75
- "Reshape",
76
- "Concat",
77
- "ReduceMean",
78
- "Pow",
79
- "Mul",
80
  "Slice",
81
  "MatMul",
82
- "Sqrt",
83
- "Transpose",
84
  "Add",
 
 
 
 
85
  "Erf",
 
 
 
 
86
  "Constant",
87
- "Softmax"
 
88
  ],
89
  "weight_type": "QInt8"
90
  },
91
  "encoder_model": {
92
  "op_types": [
93
- "Sub",
 
 
 
94
  "Gather",
95
- "Shape",
96
- "Unsqueeze",
97
  "Div",
98
  "ReduceMean",
 
99
  "Pow",
100
- "Mul",
101
  "Concat",
102
  "Reshape",
103
- "MatMul",
104
- "Transpose",
105
- "Sqrt",
106
- "Add",
107
- "Erf",
108
- "Softmax",
109
  "Constant",
110
- "Conv"
 
 
111
  ],
112
  "weight_type": "QUInt8"
113
  }
 
2
  "per_channel": false,
3
  "reduce_range": false,
4
  "per_model_config": {
5
+ "decoder_model": {
6
  "op_types": [
7
+ "Expand",
8
  "Gather",
9
+ "Squeeze",
10
+ "Range",
11
+ "Equal",
12
+ "MatMul",
13
+ "ConstantOfShape",
14
+ "Add",
15
+ "Erf",
16
+ "ReduceMean",
17
  "Reshape",
18
+ "Cast",
19
+ "Transpose",
20
  "Sqrt",
21
+ "Less",
 
 
22
  "Pow",
23
  "Concat",
24
+ "Slice",
 
 
25
  "Softmax",
26
+ "Where",
27
  "Sub",
 
28
  "Div",
29
+ "Unsqueeze",
30
+ "Shape",
31
+ "Constant",
32
+ "Mul"
 
 
 
 
 
 
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
+ "decoder_model_merged": {
37
  "op_types": [
38
+ "Expand",
39
  "Gather",
40
+ "Squeeze",
41
+ "Range",
42
+ "Equal",
43
+ "MatMul",
44
+ "ConstantOfShape",
45
+ "Add",
46
+ "Erf",
47
+ "ReduceMean",
48
  "Reshape",
49
+ "Cast",
50
+ "Transpose",
51
  "Sqrt",
52
+ "Less",
 
 
53
  "Pow",
54
  "Concat",
55
+ "Slice",
 
 
56
  "Softmax",
57
+ "Where",
58
  "Sub",
 
59
  "Div",
60
+ "If",
61
+ "Unsqueeze",
62
+ "Shape",
63
+ "Constant",
64
+ "Mul"
 
 
 
 
65
  ],
66
  "weight_type": "QInt8"
67
  },
68
  "decoder_with_past_model": {
69
  "op_types": [
70
+ "Sqrt",
 
 
 
 
 
 
 
 
 
71
  "Slice",
72
  "MatMul",
73
+ "Softmax",
74
+ "Gather",
75
  "Add",
76
+ "Sub",
77
+ "Pow",
78
+ "ReduceMean",
79
+ "Div",
80
  "Erf",
81
+ "Concat",
82
+ "Reshape",
83
+ "Unsqueeze",
84
+ "Shape",
85
  "Constant",
86
+ "Mul",
87
+ "Transpose"
88
  ],
89
  "weight_type": "QInt8"
90
  },
91
  "encoder_model": {
92
  "op_types": [
93
+ "Sqrt",
94
+ "Conv",
95
+ "MatMul",
96
+ "Softmax",
97
  "Gather",
98
+ "Add",
99
+ "Erf",
100
  "Div",
101
  "ReduceMean",
102
+ "Sub",
103
  "Pow",
 
104
  "Concat",
105
  "Reshape",
106
+ "Shape",
 
 
 
 
 
107
  "Constant",
108
+ "Mul",
109
+ "Unsqueeze",
110
+ "Transpose"
111
  ],
112
  "weight_type": "QUInt8"
113
  }