Xenova HF staff commited on
Commit
8bfeaf4
1 Parent(s): c50b99a

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 sha256:116f7dcda1e9ce5987be193cfbfefa85aeae2e80e03aca39d4602c5463046e5d
3
- size 1686022
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24a162d9454daa1ef7eadfbc5c498aa8da370fa5d6f14b9091acb758f8d4eaa3
3
+ size 1687965
onnx/decoder_model.onnx_data CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a3394a58afd855308a1fdcce769c724bfc860c0ae99b5570e8f966a40f3e7e30
3
- size 3891630080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37d8047bdced9c506618c73a628a6ffc4fe5bc22b0afb025ec0a5df6d6796e76
3
+ size 3626081280
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:70edcd042c2650a52039b9df8968f312929cd8e44ec0b0e7c9f02b5d3af643fe
3
- size 3145904
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9299eda8d2dde4396ec7de893d00c09c4d2dba6c433a41e08e0f9e1763f6ac73
3
+ size 3151825
onnx/decoder_model_merged.onnx_data CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a3394a58afd855308a1fdcce769c724bfc860c0ae99b5570e8f966a40f3e7e30
3
- size 3891630080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37d8047bdced9c506618c73a628a6ffc4fe5bc22b0afb025ec0a5df6d6796e76
3
+ size 3626081280
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2b55b394d6fcb18d270e5945aee4450ef00b8d794e3c0af24fe5a33973dc8247
3
- size 981440477
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f986f262778ead3a379ae1d275415c962116d540896406d34318cb2260803d9e
3
+ size 915058317
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6413d07da7c6ecb6b1d906c83cf1982afd0415c401d45acb2faa5399c7b61f72
3
- size 979237768
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f33d8dc7bdd76e8b2fda4f838ec001ce93e789c2e6dcf30d339f67c8dcf19b9
3
+ size 912852053
onnx/decoder_with_past_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8d417e0b0b6c7055fbf02d66068b782e7564f7476f5b0ab8a8c5e04aaa302292
3
- size 1491660
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fae8c640318528967efdb8711676c66fa3b3d3e847d8fc983cbd249a1182f53
3
+ size 1493595
onnx/decoder_with_past_model.onnx_data CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1b4fc3e889a584a43decef7ab17c43d0b06aca46af29cc65e9e4924d10210625
3
- size 3472035840
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68879149602e86721091496311b82611c9bd3101e71a35f4754a05b9487c2df2
3
+ size 3206487040
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d4631fc793b86e93f1c965a722754da5ecd7d766e5b1b0556dcf8e1990a568b8
3
- size 873894992
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76b872bba83eb2fabcedc7fb29b2b37cf3783a6ac9e8480f935a4b28cd41646d
3
+ size 807509283
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
- "Squeeze",
8
- "Cast",
9
- "Transpose",
10
  "Where",
11
- "Softmax",
12
- "Equal",
13
- "Expand",
14
  "Slice",
 
 
 
 
15
  "Unsqueeze",
16
  "Constant",
17
- "Less",
18
  "Shape",
19
- "MatMul",
20
- "Concat",
21
- "Erf",
22
- "If",
23
- "Gather",
24
- "Pow",
25
  "Sub",
 
26
  "Add",
27
  "ReduceMean",
28
  "Reshape",
29
- "ConstantOfShape",
30
- "Div",
31
- "Sqrt",
32
  "Range",
33
- "Mul"
 
 
 
 
 
 
 
34
  ],
35
  "weight_type": "QInt8"
36
  },
37
- "decoder_model": {
38
  "op_types": [
39
- "Squeeze",
40
- "Cast",
41
- "Transpose",
42
  "Where",
43
- "Softmax",
44
- "Equal",
45
- "Expand",
46
  "Slice",
 
 
 
 
 
47
  "Unsqueeze",
48
  "Constant",
49
- "Less",
50
  "Shape",
51
- "MatMul",
52
- "Concat",
53
- "Erf",
54
- "Gather",
55
- "Pow",
56
  "Sub",
 
57
  "Add",
58
  "ReduceMean",
59
  "Reshape",
60
- "ConstantOfShape",
61
- "Div",
62
- "Sqrt",
63
  "Range",
64
- "Mul"
 
 
 
 
 
 
 
65
  ],
66
  "weight_type": "QInt8"
67
  },
68
  "decoder_with_past_model": {
69
  "op_types": [
70
- "Concat",
71
- "Erf",
72
- "Gather",
73
- "Slice",
74
- "Div",
75
  "MatMul",
76
- "Pow",
77
- "Unsqueeze",
78
- "Transpose",
79
  "Sub",
 
 
80
  "Sqrt",
81
- "Softmax",
 
82
  "Add",
83
  "ReduceMean",
84
- "Constant",
85
- "Shape",
86
  "Reshape",
87
- "Mul"
 
 
 
 
 
88
  ],
89
  "weight_type": "QInt8"
90
  },
91
  "encoder_model": {
92
  "op_types": [
93
- "Concat",
94
- "Erf",
95
- "Conv",
96
- "Div",
97
- "Transpose",
98
- "Gather",
99
  "MatMul",
100
- "Pow",
101
- "Unsqueeze",
102
  "Sub",
 
 
103
  "Sqrt",
104
- "Softmax",
 
105
  "Add",
106
  "ReduceMean",
107
- "Constant",
108
- "Shape",
 
109
  "Reshape",
110
- "Mul"
 
 
 
 
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
  "Where",
8
+ "Sqrt",
9
+ "Cast",
10
+ "ConstantOfShape",
11
  "Slice",
12
+ "Erf",
13
+ "MatMul",
14
+ "Div",
15
+ "Softmax",
16
  "Unsqueeze",
17
  "Constant",
 
18
  "Shape",
 
 
 
 
 
 
19
  "Sub",
20
+ "Pow",
21
  "Add",
22
  "ReduceMean",
23
  "Reshape",
 
 
 
24
  "Range",
25
+ "Expand",
26
+ "Equal",
27
+ "Gather",
28
+ "Transpose",
29
+ "Squeeze",
30
+ "Concat",
31
+ "Mul",
32
+ "Less"
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
+ "decoder_model_merged": {
37
  "op_types": [
 
 
 
38
  "Where",
39
+ "Sqrt",
40
+ "Cast",
41
+ "ConstantOfShape",
42
  "Slice",
43
+ "Erf",
44
+ "If",
45
+ "MatMul",
46
+ "Div",
47
+ "Softmax",
48
  "Unsqueeze",
49
  "Constant",
 
50
  "Shape",
 
 
 
 
 
51
  "Sub",
52
+ "Pow",
53
  "Add",
54
  "ReduceMean",
55
  "Reshape",
 
 
 
56
  "Range",
57
+ "Expand",
58
+ "Equal",
59
+ "Gather",
60
+ "Transpose",
61
+ "Squeeze",
62
+ "Concat",
63
+ "Mul",
64
+ "Less"
65
  ],
66
  "weight_type": "QInt8"
67
  },
68
  "decoder_with_past_model": {
69
  "op_types": [
 
 
 
 
 
70
  "MatMul",
 
 
 
71
  "Sub",
72
+ "Unsqueeze",
73
+ "Gather",
74
  "Sqrt",
75
+ "Transpose",
76
+ "Pow",
77
  "Add",
78
  "ReduceMean",
79
+ "Div",
80
+ "Concat",
81
  "Reshape",
82
+ "Mul",
83
+ "Softmax",
84
+ "Slice",
85
+ "Constant",
86
+ "Erf",
87
+ "Shape"
88
  ],
89
  "weight_type": "QInt8"
90
  },
91
  "encoder_model": {
92
  "op_types": [
 
 
 
 
 
 
93
  "MatMul",
 
 
94
  "Sub",
95
+ "Unsqueeze",
96
+ "Transpose",
97
  "Sqrt",
98
+ "Gather",
99
+ "Pow",
100
  "Add",
101
  "ReduceMean",
102
+ "Div",
103
+ "Concat",
104
+ "Mul",
105
  "Reshape",
106
+ "Softmax",
107
+ "Conv",
108
+ "Constant",
109
+ "Erf",
110
+ "Shape"
111
  ],
112
  "weight_type": "QUInt8"
113
  }