Xenova HF staff commited on
Commit
d8b94eb
1 Parent(s): b71c00e

Upload folder using huggingface_hub

Browse files
config.json CHANGED
@@ -143,8 +143,7 @@
143
  50361,
144
  50362
145
  ],
146
- "torch_dtype": "float32",
147
- "transformers_version": "4.29.2",
148
  "use_cache": true,
149
  "use_weighted_layer_sum": false,
150
  "vocab_size": 51865
 
143
  50361,
144
  50362
145
  ],
146
+ "transformers_version": "4.30.2",
 
147
  "use_cache": true,
148
  "use_weighted_layer_sum": false,
149
  "vocab_size": 51865
generation_config.json CHANGED
@@ -217,5 +217,5 @@
217
  "transcribe": 50359,
218
  "translate": 50358
219
  },
220
- "transformers_version": "4.29.2"
221
  }
 
217
  "transcribe": 50359,
218
  "translate": 50358
219
  },
220
+ "transformers_version": "4.30.2"
221
  }
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:289504e89417c90b51380c46cf839d192d9a9ac94ff50d64106322636963830a
3
  size 774594491
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c09ce6b24db9c8a55eab37aba77fb217052da64fa7f30d5f0411008644bc0e15
3
  size 774594491
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d207cac77f0112c57a0d45e43ea3e38996cea95dcc49da4f229e0c8dc578531e
3
- size 197291535
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:df24179148e324c957c5740928e54c58f7894db9022949eb4ba4e5a5971aa4b9
3
+ size 196432559
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f870b64ca059126c053fe199770e71f6f69e28a871ce212c266ae4fecf8aeeef
3
- size 196598144
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48bed0449e061d9ceef7a1b6b54800e9c730478e769404309e34dca27e4a89f4
3
+ size 195739168
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bcf095f74de4ed450b92d871577848399f653d5d49cad07fec42acc333ec2c2d
3
- size 182185462
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f25f8c98570159bb7ed56df104a949ed523087a13b484f71bc9a351e50eec8b7
3
+ size 181418766
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c6bab7874f6ec1e7ee1ae0c0fe0c08705e0810f2f68eb41b354f1fa0b160d9b7
3
- size 92704740
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:736382d9bb09533d595c3c2d96e9f8716e9facfe10242b8ab4c8e802377665d3
3
+ size 92289693
quant_config.json ADDED
@@ -0,0 +1,89 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_channel": false,
3
+ "reduce_range": false,
4
+ "per_model_config": {
5
+ "encoder_model": {
6
+ "op_types": [
7
+ "Mul",
8
+ "Concat",
9
+ "Sub",
10
+ "ReduceMean",
11
+ "Softmax",
12
+ "Transpose",
13
+ "Div",
14
+ "Pow",
15
+ "Gather",
16
+ "Constant",
17
+ "Conv",
18
+ "Erf",
19
+ "MatMul",
20
+ "Unsqueeze",
21
+ "Sqrt",
22
+ "Add",
23
+ "Shape",
24
+ "Reshape"
25
+ ],
26
+ "weight_type": "QUInt8"
27
+ },
28
+ "decoder_model": {
29
+ "op_types": [
30
+ "Range",
31
+ "ConstantOfShape",
32
+ "MatMul",
33
+ "Sqrt",
34
+ "Mul",
35
+ "Concat",
36
+ "Less",
37
+ "ReduceMean",
38
+ "Sub",
39
+ "Pow",
40
+ "Erf",
41
+ "Slice",
42
+ "Cast",
43
+ "Gather",
44
+ "Where",
45
+ "Softmax",
46
+ "Constant",
47
+ "Squeeze",
48
+ "Add",
49
+ "Shape",
50
+ "Reshape",
51
+ "Transpose",
52
+ "Div",
53
+ "Unsqueeze",
54
+ "Expand",
55
+ "Equal"
56
+ ],
57
+ "weight_type": "QInt8"
58
+ },
59
+ "decoder_model_merged": {
60
+ "op_types": [
61
+ "If"
62
+ ],
63
+ "weight_type": "QInt8"
64
+ },
65
+ "decoder_with_past_model": {
66
+ "op_types": [
67
+ "Mul",
68
+ "Concat",
69
+ "Sub",
70
+ "ReduceMean",
71
+ "Softmax",
72
+ "Transpose",
73
+ "Div",
74
+ "Pow",
75
+ "Shape",
76
+ "Constant",
77
+ "Erf",
78
+ "Slice",
79
+ "MatMul",
80
+ "Unsqueeze",
81
+ "Sqrt",
82
+ "Add",
83
+ "Gather",
84
+ "Reshape"
85
+ ],
86
+ "weight_type": "QInt8"
87
+ }
88
+ }
89
+ }