Xenova HF staff commited on
Commit
4585a90
1 Parent(s): f6c0672

Upload folder using huggingface_hub

Browse files
config.json CHANGED
@@ -54,7 +54,7 @@
54
  "scale_embedding": true,
55
  "share_encoder_decoder_embeddings": true,
56
  "static_position_embeddings": true,
57
- "transformers_version": "4.32.0.dev0",
58
  "use_cache": true,
59
  "vocab_size": 65001
60
  }
 
54
  "scale_embedding": true,
55
  "share_encoder_decoder_embeddings": true,
56
  "static_position_embeddings": true,
57
+ "transformers_version": "4.34.0.dev0",
58
  "use_cache": true,
59
  "vocab_size": 65001
60
  }
generation_config.json CHANGED
@@ -1,5 +1,4 @@
1
  {
2
- "_from_model_config": true,
3
  "bad_words_ids": [
4
  [
5
  65000
@@ -12,5 +11,6 @@
12
  "max_length": 512,
13
  "num_beams": 6,
14
  "pad_token_id": 65000,
15
- "transformers_version": "4.32.0.dev0"
 
16
  }
 
1
  {
 
2
  "bad_words_ids": [
3
  [
4
  65000
 
11
  "max_length": 512,
12
  "num_beams": 6,
13
  "pad_token_id": 65000,
14
+ "renormalize_logits": true,
15
+ "transformers_version": "4.34.0.dev0"
16
  }
onnx/decoder_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d6005cf82395d5981f33b856afe794f13ab0f31a3c696d4c0cf70ae2db869c31
3
- size 368725660
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d07d3c7daa90c0d425b8f4a53fe8f221a9b071b2dac5d76a0c4eae52f7a018ee
3
+ size 235603669
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9f522bb11361e56d5c77b156b794a5e2c123f1a5a1eb69e0db8e1d21d399f246
3
- size 368961142
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b1423b06499b78c8c2856468bf98aaab8bf6d7497d9902227186031f6fb289b
3
+ size 235839236
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8133e924cac321c334ca99b69b34d5bc08ffebe929b24244545d57330b32b0f3
3
- size 93819732
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e131732653a6d3cec42a8b65f8844980042b036bdaca931446bc4b3530bc3506
3
+ size 60212803
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:338df537cff1d7a8546c5a99374cf6485c63740831d770448fc3ec6294806ae5
3
- size 93448376
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c961ea33a62ea84e73c9ec94c0737e65d06b5f73a7d0398b018fbbb7262d929
3
+ size 59842101
onnx/decoder_with_past_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3b52aa52401399c0a40f2bb665a6b8ce83c62c096f838042d5c6661dbb33d585
3
- size 356075151
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fed67a95b9bf291f646aa2a17f0f64488a5422feb14c38e89fe0e77a3b1c0eaa
3
+ size 222953160
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:37a75b0b2aaa19d87f5856f6b2b26001e3c8f0a14317cf376438687f29b9be2d
3
- size 90176912
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:196bd1705ae0eac256f1a9ab8c82c243decdad124c5ace67fe073d9e0fc6956e
3
+ size 56570637
quantize_config.json CHANGED
@@ -2,122 +2,122 @@
2
  "per_channel": true,
3
  "reduce_range": true,
4
  "per_model_config": {
5
- "encoder_model": {
6
  "op_types": [
7
- "ReduceMean",
8
- "Expand",
9
- "Unsqueeze",
10
  "Softmax",
11
- "Where",
12
- "Pow",
13
- "Equal",
14
- "ConstantOfShape",
15
- "Shape",
16
  "Div",
17
- "Reshape",
 
 
 
 
 
18
  "Sub",
19
- "Cast",
20
- "Concat",
21
- "Constant",
22
  "Mul",
23
- "Sigmoid",
24
  "Gather",
25
- "Sqrt",
 
 
 
26
  "Transpose",
27
- "MatMul",
 
 
28
  "Range",
29
- "Add"
 
30
  ],
31
  "weight_type": "QInt8"
32
  },
33
- "decoder_with_past_model": {
34
  "op_types": [
35
- "ReduceMean",
36
- "Expand",
37
- "Unsqueeze",
38
  "Softmax",
39
- "Where",
40
- "Pow",
41
- "Equal",
42
- "ConstantOfShape",
43
- "Shape",
44
  "Div",
45
- "Reshape",
 
 
 
46
  "Sub",
47
- "Cast",
48
- "Concat",
49
- "Constant",
50
  "Mul",
51
- "Sigmoid",
52
  "Gather",
53
- "Sqrt",
 
 
 
54
  "Transpose",
55
- "MatMul",
 
56
  "Range",
57
- "Add"
 
58
  ],
59
  "weight_type": "QInt8"
60
  },
61
- "decoder_model": {
62
  "op_types": [
63
- "ReduceMean",
64
- "Expand",
65
- "Squeeze",
66
- "Unsqueeze",
67
  "Softmax",
68
- "Where",
69
- "Less",
70
- "Pow",
71
- "Equal",
72
- "ConstantOfShape",
73
- "Shape",
74
  "Div",
75
- "Reshape",
 
 
 
 
 
76
  "Sub",
77
- "Cast",
78
- "Concat",
79
- "Constant",
80
  "Mul",
81
- "Sigmoid",
82
  "Gather",
83
- "Sqrt",
84
- "Slice",
 
 
85
  "Transpose",
86
- "MatMul",
 
 
87
  "Range",
88
- "Add"
 
89
  ],
90
  "weight_type": "QInt8"
91
  },
92
- "decoder_model_merged": {
93
  "op_types": [
94
- "ReduceMean",
95
- "Expand",
96
- "Squeeze",
97
- "Unsqueeze",
98
  "Softmax",
99
- "If",
100
- "Less",
101
- "Where",
102
- "Equal",
103
- "Pow",
104
- "ConstantOfShape",
105
- "Shape",
106
  "Div",
107
- "Reshape",
 
 
 
108
  "Sub",
109
- "Cast",
110
- "Concat",
111
- "Constant",
112
  "Mul",
113
- "Sigmoid",
114
  "Gather",
115
- "Sqrt",
116
- "Slice",
 
 
117
  "Transpose",
118
- "MatMul",
 
119
  "Range",
120
- "Add"
 
121
  ],
122
  "weight_type": "QInt8"
123
  }
 
2
  "per_channel": true,
3
  "reduce_range": true,
4
  "per_model_config": {
5
+ "decoder_model": {
6
  "op_types": [
7
+ "Constant",
8
+ "Reshape",
 
9
  "Softmax",
10
+ "Slice",
 
 
 
 
11
  "Div",
12
+ "Sigmoid",
13
+ "Squeeze",
14
+ "Unsqueeze",
15
+ "MatMul",
16
+ "Add",
17
+ "ReduceMean",
18
  "Sub",
19
+ "Shape",
 
 
20
  "Mul",
 
21
  "Gather",
22
+ "Expand",
23
+ "Equal",
24
+ "Pow",
25
+ "Concat",
26
  "Transpose",
27
+ "Less",
28
+ "Sqrt",
29
+ "Where",
30
  "Range",
31
+ "Cast",
32
+ "ConstantOfShape"
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
+ "encoder_model": {
37
  "op_types": [
38
+ "Constant",
39
+ "Reshape",
 
40
  "Softmax",
41
+ "Sigmoid",
 
 
 
 
42
  "Div",
43
+ "Unsqueeze",
44
+ "MatMul",
45
+ "Add",
46
+ "ReduceMean",
47
  "Sub",
48
+ "Shape",
 
 
49
  "Mul",
 
50
  "Gather",
51
+ "Expand",
52
+ "Equal",
53
+ "Pow",
54
+ "Concat",
55
  "Transpose",
56
+ "Sqrt",
57
+ "Where",
58
  "Range",
59
+ "Cast",
60
+ "ConstantOfShape"
61
  ],
62
  "weight_type": "QInt8"
63
  },
64
+ "decoder_model_merged": {
65
  "op_types": [
66
+ "Constant",
67
+ "Reshape",
68
+ "If",
 
69
  "Softmax",
70
+ "Slice",
 
 
 
 
 
71
  "Div",
72
+ "Sigmoid",
73
+ "Squeeze",
74
+ "Unsqueeze",
75
+ "MatMul",
76
+ "Add",
77
+ "ReduceMean",
78
  "Sub",
79
+ "Shape",
 
 
80
  "Mul",
 
81
  "Gather",
82
+ "Expand",
83
+ "Equal",
84
+ "Pow",
85
+ "Concat",
86
  "Transpose",
87
+ "Less",
88
+ "Sqrt",
89
+ "Where",
90
  "Range",
91
+ "Cast",
92
+ "ConstantOfShape"
93
  ],
94
  "weight_type": "QInt8"
95
  },
96
+ "decoder_with_past_model": {
97
  "op_types": [
98
+ "Constant",
99
+ "Reshape",
 
 
100
  "Softmax",
101
+ "Sigmoid",
 
 
 
 
 
 
102
  "Div",
103
+ "Unsqueeze",
104
+ "MatMul",
105
+ "Add",
106
+ "ReduceMean",
107
  "Sub",
108
+ "Shape",
 
 
109
  "Mul",
 
110
  "Gather",
111
+ "Expand",
112
+ "Equal",
113
+ "Pow",
114
+ "Concat",
115
  "Transpose",
116
+ "Sqrt",
117
+ "Where",
118
  "Range",
119
+ "Cast",
120
+ "ConstantOfShape"
121
  ],
122
  "weight_type": "QInt8"
123
  }
tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff