Xenova HF staff commited on
Commit
fc805ad
·
1 Parent(s): 5b3aabe

Upload folder using huggingface_hub

Browse files
config.json CHANGED
@@ -55,7 +55,7 @@
55
  "scale_embedding": true,
56
  "share_encoder_decoder_embeddings": true,
57
  "static_position_embeddings": true,
58
- "transformers_version": "4.32.0.dev0",
59
  "use_cache": true,
60
  "vocab_size": 65001
61
  }
 
55
  "scale_embedding": true,
56
  "share_encoder_decoder_embeddings": true,
57
  "static_position_embeddings": true,
58
+ "transformers_version": "4.34.0.dev0",
59
  "use_cache": true,
60
  "vocab_size": 65001
61
  }
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": 4,
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": 4,
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 sha256:694f62d4833694bf2cff7222797d770ee42e1437e2579f6e4986c7546250e420
3
- size 368725660
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6d23b324c9c871a16e1d21c5a54dab9b48c22c919771bd53bb107b35f4fa097
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:b801f40a1713d9845a21ff04c4eef10732f8b9d7b521b587efd166ef27fc13e9
3
- size 368961142
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6d2bb9fc704b42b8bf8e6e67ee6b4e865cab674706b29a10e5996ac135e5d625
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:caedc065406da961a4c9471dbb5d2a227b342cb468966b9144a905d7c39dd4f7
3
- size 93819733
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4cd91ab30240d295ce907b5100826031838c672005e57e524d711122d75605fa
3
+ size 60212804
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f027e5cf3046ee06186ac75955a2c8f4b2eae4376d66f91a6bf10e5a0f26787b
3
- size 93448377
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d518f1f0041b2831f0aeb5a087cf899742cd9d4b574f6b98ad9f816b8c06925b
3
+ size 59842102
onnx/decoder_with_past_model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e94deb365cfb273cba9cc1949ebbc00b4e630bea3621cd75baa7ec201cef01be
3
- size 356075151
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d8816abe35a5ff6cfcbc62444228703b1e2ee196542cf42d75ef65666e13290
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 sha256:58c687cf25cd3ea248e3da98979d53157e5d0303a927122f11d82c520f6af670
3
- size 90176913
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c3d587fc040d6d9669faa45658d51c241aea89e8c25617814d136426aed7c11f
3
+ size 56570638
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
- "Reshape",
8
- "Where",
9
- "Add",
10
  "Sub",
11
- "Transpose",
12
- "ConstantOfShape",
13
- "Gather",
14
- "Constant",
15
  "Expand",
16
- "Softmax",
17
- "Range",
18
- "Shape",
19
- "Concat",
20
  "Div",
 
 
21
  "Cast",
22
- "MatMul",
23
- "Sigmoid",
24
- "Unsqueeze",
25
  "Equal",
26
- "Mul",
27
- "ReduceMean",
 
 
 
 
28
  "Sqrt",
29
- "Pow"
 
 
 
 
 
 
 
 
 
 
30
  ],
31
  "weight_type": "QInt8"
32
  },
33
- "decoder_with_past_model": {
34
  "op_types": [
35
- "Reshape",
36
- "Where",
37
- "Add",
38
  "Sub",
39
- "Transpose",
40
- "ConstantOfShape",
41
- "Gather",
42
- "Constant",
43
  "Expand",
44
- "Softmax",
45
- "Range",
46
- "Shape",
47
- "Concat",
48
  "Div",
 
 
49
  "Cast",
50
- "MatMul",
51
- "Sigmoid",
52
- "Unsqueeze",
53
  "Equal",
54
- "Mul",
55
- "ReduceMean",
 
 
56
  "Sqrt",
57
- "Pow"
 
 
 
 
 
 
 
 
 
 
58
  ],
59
  "weight_type": "QInt8"
60
  },
61
- "decoder_model": {
62
  "op_types": [
63
- "Reshape",
64
- "Where",
65
- "Slice",
66
- "Add",
67
  "Sub",
68
- "Transpose",
 
 
69
  "ConstantOfShape",
 
 
 
 
 
70
  "Gather",
71
- "Constant",
72
- "Expand",
73
  "Squeeze",
74
- "Softmax",
 
 
75
  "Range",
76
- "Shape",
77
- "Less",
78
- "Concat",
79
- "Div",
80
- "Cast",
81
  "MatMul",
82
- "Sigmoid",
83
- "Unsqueeze",
84
- "Equal",
85
- "Mul",
86
  "ReduceMean",
87
- "Sqrt",
88
- "Pow"
 
 
 
 
 
89
  ],
90
  "weight_type": "QInt8"
91
  },
92
- "decoder_model_merged": {
93
  "op_types": [
94
- "Reshape",
95
- "Where",
96
- "Slice",
97
- "Add",
98
  "Sub",
99
- "Transpose",
100
- "ConstantOfShape",
101
- "Gather",
102
- "Constant",
103
  "Expand",
104
- "Squeeze",
105
- "Softmax",
106
- "Range",
107
- "Shape",
108
- "Less",
109
- "Concat",
110
  "Div",
 
 
 
111
  "Cast",
112
- "MatMul",
113
  "Sigmoid",
114
- "Unsqueeze",
115
- "Equal",
116
- "If",
117
- "Mul",
118
- "ReduceMean",
119
  "Sqrt",
120
- "Pow"
 
 
 
 
 
 
 
 
 
 
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
  "Sub",
8
+ "Slice",
 
 
 
9
  "Expand",
 
 
 
 
10
  "Div",
11
+ "ConstantOfShape",
12
+ "Add",
13
  "Cast",
 
 
 
14
  "Equal",
15
+ "Shape",
16
+ "Less",
17
+ "Gather",
18
+ "Sigmoid",
19
+ "Transpose",
20
+ "Squeeze",
21
  "Sqrt",
22
+ "Pow",
23
+ "Range",
24
+ "MatMul",
25
+ "ReduceMean",
26
+ "Constant",
27
+ "Where",
28
+ "Concat",
29
+ "Softmax",
30
+ "Mul",
31
+ "Reshape",
32
+ "Unsqueeze"
33
  ],
34
  "weight_type": "QInt8"
35
  },
36
+ "encoder_model": {
37
  "op_types": [
 
 
 
38
  "Sub",
 
 
 
 
39
  "Expand",
 
 
 
 
40
  "Div",
41
+ "ConstantOfShape",
42
+ "Add",
43
  "Cast",
 
 
 
44
  "Equal",
45
+ "Shape",
46
+ "Sigmoid",
47
+ "Gather",
48
+ "Transpose",
49
  "Sqrt",
50
+ "Pow",
51
+ "Range",
52
+ "MatMul",
53
+ "ReduceMean",
54
+ "Constant",
55
+ "Where",
56
+ "Concat",
57
+ "Softmax",
58
+ "Mul",
59
+ "Reshape",
60
+ "Unsqueeze"
61
  ],
62
  "weight_type": "QInt8"
63
  },
64
+ "decoder_model_merged": {
65
  "op_types": [
 
 
 
 
66
  "Sub",
67
+ "Slice",
68
+ "Expand",
69
+ "Div",
70
  "ConstantOfShape",
71
+ "Add",
72
+ "Cast",
73
+ "Equal",
74
+ "Shape",
75
+ "Less",
76
  "Gather",
77
+ "Sigmoid",
78
+ "Transpose",
79
  "Squeeze",
80
+ "Sqrt",
81
+ "If",
82
+ "Pow",
83
  "Range",
 
 
 
 
 
84
  "MatMul",
 
 
 
 
85
  "ReduceMean",
86
+ "Constant",
87
+ "Where",
88
+ "Concat",
89
+ "Softmax",
90
+ "Mul",
91
+ "Reshape",
92
+ "Unsqueeze"
93
  ],
94
  "weight_type": "QInt8"
95
  },
96
+ "decoder_with_past_model": {
97
  "op_types": [
 
 
 
 
98
  "Sub",
 
 
 
 
99
  "Expand",
 
 
 
 
 
 
100
  "Div",
101
+ "ConstantOfShape",
102
+ "Add",
103
+ "Equal",
104
  "Cast",
105
+ "Shape",
106
  "Sigmoid",
107
+ "Gather",
108
+ "Transpose",
 
 
 
109
  "Sqrt",
110
+ "Pow",
111
+ "Range",
112
+ "MatMul",
113
+ "ReduceMean",
114
+ "Constant",
115
+ "Where",
116
+ "Concat",
117
+ "Softmax",
118
+ "Mul",
119
+ "Reshape",
120
+ "Unsqueeze"
121
  ],
122
  "weight_type": "QInt8"
123
  }
tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff