Xenova HF staff commited on
Commit
abbbb97
1 Parent(s): 49e3862

Upload folder using huggingface_hub

Browse files
config.json CHANGED
@@ -141,8 +141,7 @@
141
  50361,
142
  50362
143
  ],
144
- "torch_dtype": "float32",
145
- "transformers_version": "4.29.2",
146
  "use_cache": true,
147
  "use_weighted_layer_sum": false,
148
  "vocab_size": 51865
 
141
  50361,
142
  50362
143
  ],
144
+ "transformers_version": "4.30.2",
 
145
  "use_cache": true,
146
  "use_weighted_layer_sum": false,
147
  "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:eb77e2680627baa157fbba1e95fe6a28610420814d2685ee78bccf0357de44e0
3
  size 2761864
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34eb70fe697d78afa2517f36b32933ab1d7d19e76897d3c8d8156f93dffd3ad6
3
  size 2761864
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:75b7dfbbaa56176ea2bac41a3a7ecefd72990da0c1d3c90f75ef836a5580e913
3
- size 983867549
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:250fc0db44d671af75069b8b5d3c2fb001503dc7a78cd5a93655eb237053a0c0
3
+ size 980944149
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d5a95c979c9dc944cb7246c17f4c7a36d196436dd495442eaf8d0c5165f1dba2
3
- size 981928446
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bff22b6a6dee56fb9da83164265061a5e2f394be1b182d3b5af17778e5dbd4a9
3
+ size 979005046
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:af7524a162c21c2dfbae624df361effd28538f83513d15cf5d01cb9196249e6c
3
- size 876151256
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:792e3ec0ed3737b6807ff7b53a0c2005c089a7c73982789186a4a5b6766b6be9
3
+ size 873637776
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1ab631c21615158bfda3a824393ab3823dbacf4f204764e3849894f6cc0e132d
3
- size 646822577
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d25cdce63acc9b2ca2de12091dfdc7977e2bdfff75a74a170e1a710e0706deb
3
+ size 644978452
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
+ "Erf",
8
+ "Sub",
9
+ "Pow",
10
+ "Reshape",
11
+ "ReduceMean",
12
+ "MatMul",
13
+ "Constant",
14
+ "Conv",
15
+ "Div",
16
+ "Gather",
17
+ "Mul",
18
+ "Unsqueeze",
19
+ "Concat",
20
+ "Shape",
21
+ "Sqrt",
22
+ "Transpose",
23
+ "Softmax",
24
+ "Add"
25
+ ],
26
+ "weight_type": "QUInt8"
27
+ },
28
+ "decoder_model": {
29
+ "op_types": [
30
+ "MatMul",
31
+ "Constant",
32
+ "Gather",
33
+ "Concat",
34
+ "Cast",
35
+ "Squeeze",
36
+ "Softmax",
37
+ "Range",
38
+ "Equal",
39
+ "Mul",
40
+ "Less",
41
+ "Erf",
42
+ "Reshape",
43
+ "Pow",
44
+ "Div",
45
+ "Expand",
46
+ "Slice",
47
+ "Add",
48
+ "ConstantOfShape",
49
+ "Sub",
50
+ "ReduceMean",
51
+ "Where",
52
+ "Unsqueeze",
53
+ "Shape",
54
+ "Sqrt",
55
+ "Transpose"
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
+ "Erf",
68
+ "Reshape",
69
+ "Sub",
70
+ "Pow",
71
+ "ReduceMean",
72
+ "MatMul",
73
+ "Constant",
74
+ "Gather",
75
+ "Div",
76
+ "Unsqueeze",
77
+ "Concat",
78
+ "Slice",
79
+ "Mul",
80
+ "Shape",
81
+ "Sqrt",
82
+ "Transpose",
83
+ "Softmax",
84
+ "Add"
85
+ ],
86
+ "weight_type": "QInt8"
87
+ }
88
+ }
89
+ }