Xenova HF staff commited on
Commit
b645993
1 Parent(s): 9cc962f

Upload folder using huggingface_hub

Browse files
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:55fbbfbdd774d0f4cc2a92f59f11e58584539d3a92dfb6536f65d4d3c1c4f57a
3
  size 791453859
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85b04628082c655fe5a8cfee14893b0f03cc3e662d5352320aa401fe75aafc15
3
  size 791453859
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4699d4bb1950bb65f77e2f6ce1b26a51f26fdb5813e22581efcf0c40fd8b95b4
3
- size 300505575
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5675c823efffcc4da240db6f8ebcd0d5c5effa2038dcb52cade133b6d93227cd
3
+ size 299954775
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3a960ff52fcdff888c99beb87c7a497165a3e2201b3c24777ce295b37c546062
3
- size 232856903
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6889a8fdc9daf18b7aa3c6124684eaffd0fb2fba6ee0103064ad719d79d7bf0a
3
+ size 232306103
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c964f2c88afb32affe47ab3d531c48617ebe071b393ad5eec09c1e410948b4bf
3
- size 232865860
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66061e6a38529a523db921f29bebea1086e4753e60daff773ecfe8fb48ad0b10
3
+ size 232315060
quantize_config.json CHANGED
@@ -1,98 +1,98 @@
1
  {
2
- "per_channel": true,
3
- "reduce_range": true,
4
  "per_model_config": {
5
  "decoder_model": {
6
  "op_types": [
7
- "Transpose",
8
- "ReduceMean",
 
 
 
9
  "Div",
10
- "Gather",
11
  "And",
 
 
 
 
12
  "Add",
13
- "Constant",
14
- "Mul",
15
  "Concat",
16
- "Sub",
17
- "Split",
18
- "Pow",
19
- "Tanh",
20
- "ConstantOfShape",
21
- "Softmax",
22
  "Cast",
23
- "Sqrt",
24
- "Equal",
25
- "Squeeze",
26
- "Shape",
27
  "MatMul",
28
  "Where",
 
29
  "Reshape",
30
- "Unsqueeze",
31
- "Slice",
32
- "CumSum"
 
 
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
  "decoder_model_merged": {
37
  "op_types": [
38
- "Transpose",
39
- "ReduceMean",
 
 
 
40
  "Div",
41
- "Gather",
42
  "And",
 
 
 
 
43
  "Add",
44
- "Constant",
45
- "Mul",
46
  "Concat",
47
- "If",
48
- "Sub",
49
- "Split",
50
- "Pow",
51
- "Tanh",
52
- "ConstantOfShape",
53
- "Softmax",
54
  "Cast",
55
- "Sqrt",
56
- "Equal",
57
- "Squeeze",
58
- "Shape",
59
  "MatMul",
60
  "Where",
 
61
  "Reshape",
62
- "Unsqueeze",
63
- "Slice",
64
- "CumSum"
 
 
 
65
  ],
66
  "weight_type": "QInt8"
67
  },
68
  "decoder_with_past_model": {
69
  "op_types": [
70
- "Transpose",
71
- "ReduceMean",
 
 
 
72
  "Div",
73
- "Gather",
74
  "And",
 
 
 
 
75
  "Add",
76
- "Constant",
77
- "Mul",
78
  "Concat",
79
- "Sub",
80
- "Split",
81
- "Pow",
82
- "Tanh",
83
- "ConstantOfShape",
84
- "Softmax",
85
  "Cast",
86
- "Sqrt",
87
- "Equal",
88
- "Squeeze",
89
- "Shape",
90
  "MatMul",
91
  "Where",
 
92
  "Reshape",
93
- "Unsqueeze",
94
- "Slice",
95
- "CumSum"
 
 
96
  ],
97
  "weight_type": "QInt8"
98
  }
 
1
  {
2
+ "per_channel": false,
3
+ "reduce_range": false,
4
  "per_model_config": {
5
  "decoder_model": {
6
  "op_types": [
7
+ "Tanh",
8
+ "Softmax",
9
+ "Unsqueeze",
10
+ "Equal",
11
+ "Mul",
12
  "Div",
13
+ "Pow",
14
  "And",
15
+ "Sub",
16
+ "Shape",
17
+ "CumSum",
18
+ "Transpose",
19
  "Add",
20
+ "Slice",
 
21
  "Concat",
 
 
 
 
 
 
22
  "Cast",
23
+ "ReduceMean",
 
 
 
24
  "MatMul",
25
  "Where",
26
+ "Squeeze",
27
  "Reshape",
28
+ "Constant",
29
+ "ConstantOfShape",
30
+ "Sqrt",
31
+ "Split",
32
+ "Gather"
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
  "decoder_model_merged": {
37
  "op_types": [
38
+ "Tanh",
39
+ "Softmax",
40
+ "Unsqueeze",
41
+ "Equal",
42
+ "Mul",
43
  "Div",
44
+ "Pow",
45
  "And",
46
+ "Sub",
47
+ "Shape",
48
+ "CumSum",
49
+ "Transpose",
50
  "Add",
51
+ "Slice",
 
52
  "Concat",
 
 
 
 
 
 
 
53
  "Cast",
54
+ "ReduceMean",
 
 
 
55
  "MatMul",
56
  "Where",
57
+ "Squeeze",
58
  "Reshape",
59
+ "Constant",
60
+ "ConstantOfShape",
61
+ "Sqrt",
62
+ "If",
63
+ "Split",
64
+ "Gather"
65
  ],
66
  "weight_type": "QInt8"
67
  },
68
  "decoder_with_past_model": {
69
  "op_types": [
70
+ "Tanh",
71
+ "Softmax",
72
+ "Unsqueeze",
73
+ "Equal",
74
+ "Mul",
75
  "Div",
76
+ "Pow",
77
  "And",
78
+ "Sub",
79
+ "Shape",
80
+ "CumSum",
81
+ "Transpose",
82
  "Add",
83
+ "Slice",
 
84
  "Concat",
 
 
 
 
 
 
85
  "Cast",
86
+ "ReduceMean",
 
 
 
87
  "MatMul",
88
  "Where",
89
+ "Squeeze",
90
  "Reshape",
91
+ "Constant",
92
+ "ConstantOfShape",
93
+ "Sqrt",
94
+ "Split",
95
+ "Gather"
96
  ],
97
  "weight_type": "QInt8"
98
  }