Xenova HF staff commited on
Commit
d125a8a
1 Parent(s): 1eff9d1

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
@@ -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:eef3d2a0334f304e241c1cc0b222b0a47052c3563b1d717dbf9d5227b9235867
3
  size 314710581
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17eeff2508a474685cec5b46587c13eb7d0d28f921ce64d08b8ef3e7a9c2938e
3
  size 314710581
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fc5af04192733a84ff10daebcdc47de39bb91c68c679ad5aa0f13ae4565b113e
3
- size 80632020
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:edccc6c79bf0f9b46a99770172cf3160154734640795f84c6fc516932b7b87b9
3
+ size 80172704
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b0feb436e0ec14bf2f277bcbd6c198cc996e4535997cc3f14c36f18baa7399d9
3
- size 80282184
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0302f2ff47b04e6715e5812a1313084d404cbdeff4a55fdab22e4c5f7c00aa2
3
+ size 79822868
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:37806369274d0c13448c8024571d4be15c28a6adcda9887655f222d518206ad3
3
- size 77025486
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54d7fad0e7189431eab05cc42fb488f7d8b8011070c70e49aa9d833cb4a917c2
3
+ size 76596950
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:589ca79a883dd2dfae23acd7f801df018d3ffef9ed3eaa63eecfacff6bd18a6d
3
- size 23321703
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e360d86928a0232eb6100f2a36fa4a358c101e0a6a39763e5eefe139dca5fe71
3
+ size 23183300
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
+ "Pow",
8
+ "Gather",
9
+ "Mul",
10
+ "ReduceMean",
11
+ "Conv",
12
+ "Erf",
13
+ "MatMul",
14
+ "Concat",
15
+ "Sqrt",
16
+ "Add",
17
+ "Constant",
18
+ "Sub",
19
+ "Softmax",
20
+ "Reshape",
21
+ "Unsqueeze",
22
+ "Transpose",
23
+ "Shape",
24
+ "Div"
25
+ ],
26
+ "weight_type": "QUInt8"
27
+ },
28
+ "decoder_model": {
29
+ "op_types": [
30
+ "Pow",
31
+ "Expand",
32
+ "Concat",
33
+ "Less",
34
+ "Sqrt",
35
+ "Add",
36
+ "Unsqueeze",
37
+ "Transpose",
38
+ "Shape",
39
+ "Gather",
40
+ "Mul",
41
+ "ReduceMean",
42
+ "Erf",
43
+ "MatMul",
44
+ "ConstantOfShape",
45
+ "Constant",
46
+ "Softmax",
47
+ "Range",
48
+ "Reshape",
49
+ "Cast",
50
+ "Div",
51
+ "Where",
52
+ "Slice",
53
+ "Sub",
54
+ "Equal",
55
+ "Squeeze"
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
+ "Gather",
68
+ "Pow",
69
+ "Mul",
70
+ "ReduceMean",
71
+ "Concat",
72
+ "Slice",
73
+ "MatMul",
74
+ "Erf",
75
+ "Sqrt",
76
+ "Add",
77
+ "Constant",
78
+ "Sub",
79
+ "Softmax",
80
+ "Reshape",
81
+ "Unsqueeze",
82
+ "Transpose",
83
+ "Shape",
84
+ "Div"
85
+ ],
86
+ "weight_type": "QInt8"
87
+ }
88
+ }
89
+ }