Xenova HF staff commited on
Commit
180390a
1 Parent(s): e7d9cf6

Upload folder using huggingface_hub

Browse files
config.json CHANGED
@@ -145,8 +145,7 @@
145
  50361,
146
  50362
147
  ],
148
- "torch_dtype": "float32",
149
- "transformers_version": "4.29.2",
150
  "use_cache": true,
151
  "use_weighted_layer_sum": false,
152
  "vocab_size": 51865
 
145
  50361,
146
  50362
147
  ],
148
+ "transformers_version": "4.30.2",
 
149
  "use_cache": true,
150
  "use_weighted_layer_sum": false,
151
  "vocab_size": 51865
generation_config.json CHANGED
@@ -216,5 +216,5 @@
216
  "transcribe": 50359,
217
  "translate": 50358
218
  },
219
- "transformers_version": "4.29.2"
220
  }
 
216
  "transcribe": 50359,
217
  "translate": 50358
218
  },
219
+ "transformers_version": "4.30.2"
220
  }
onnx/decoder_model_merged.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c00671545310903db57ce48f52ec5d987345cc5fc7d8dc0e5e878c0cd0a690f2
3
  size 1926573
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f82a91e07ab23f138384c1d6aba1f20b3437ccb6beb3eaa21b4ef74cdc292ed
3
  size 1926573
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6c1ad5fdbcaa035099dfde799684204d373cbb0881f1d9618b21a1f158316405
3
- size 517265480
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0bdb36b4facd9a224dde4533588f876ab0fb3599001bde4869a39e014594ce13
3
+ size 515407456
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:87c69b1d58197c44ab1959c8e8ea0138786a112c38e042a4a072717f51af64ee
3
- size 515877306
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf74fbc6d746531b96673d35e8bd455bb049a4f570847b925e9685fe426bc5a0
3
+ size 514019282
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:09aa8d19794ec32742feecd63248bbe732cfba54136bcea917ba29a21f140463
3
- size 464952727
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e81dd3275a51a8fd1a41247b7cfb4bba5df7d10b3814640c7920579bcca19ba6
3
+ size 463340703
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:75d4b610e4852847ad7bca994733d4ae0526853d240ee51aed631817410e6832
3
- size 314503795
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d01068c82a2d3e963ffe9047b5cb802539a4f26980434c00f844d431a794ff1
3
+ size 313397176
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
+ "Sqrt",
8
+ "Conv",
9
+ "Shape",
10
+ "Gather",
11
+ "Transpose",
12
+ "ReduceMean",
13
+ "Constant",
14
+ "Mul",
15
+ "Sub",
16
+ "Add",
17
+ "Erf",
18
+ "Concat",
19
+ "Softmax",
20
+ "Div",
21
+ "Pow",
22
+ "MatMul",
23
+ "Unsqueeze",
24
+ "Reshape"
25
+ ],
26
+ "weight_type": "QUInt8"
27
+ },
28
+ "decoder_model": {
29
+ "op_types": [
30
+ "Gather",
31
+ "Expand",
32
+ "Slice",
33
+ "ReduceMean",
34
+ "Unsqueeze",
35
+ "Pow",
36
+ "Cast",
37
+ "Sqrt",
38
+ "Equal",
39
+ "Range",
40
+ "Erf",
41
+ "Reshape",
42
+ "Shape",
43
+ "ConstantOfShape",
44
+ "Transpose",
45
+ "Constant",
46
+ "Less",
47
+ "Concat",
48
+ "Add",
49
+ "Div",
50
+ "Squeeze",
51
+ "Where",
52
+ "Mul",
53
+ "Sub",
54
+ "Softmax",
55
+ "MatMul"
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
+ "Sqrt",
68
+ "Gather",
69
+ "Shape",
70
+ "Transpose",
71
+ "Slice",
72
+ "ReduceMean",
73
+ "Constant",
74
+ "Mul",
75
+ "Sub",
76
+ "Concat",
77
+ "Add",
78
+ "Softmax",
79
+ "Erf",
80
+ "Unsqueeze",
81
+ "Pow",
82
+ "Div",
83
+ "MatMul",
84
+ "Reshape"
85
+ ],
86
+ "weight_type": "QInt8"
87
+ }
88
+ }
89
+ }