Xenova HF staff commited on
Commit
4476b95
1 Parent(s): 35aa69d

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:ca286b0628b2bf1a25b8f7d36640926d3394b24b81c6dfbd1e380920d72d92c1
3
  size 198197535
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a3f5dfc9cf89e4910a1be6574ed519181eefc743fd125ddcd4a97632fd44e94
3
  size 198197535
onnx/decoder_model_merged_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a5e98c24f94eeefadebec026324a67baec6ff3ea504bb4f140134b95d8eb6d00
3
- size 50943703
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3630e9e28169f8664cdaed6d2b0f00abae079bae7e4336c249308022635721b4
3
+ size 50584327
onnx/decoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c2741fcc4bf68bebccf3bd5c607a0c541d8ab0938444d6deb0f979ebfbed4367
3
- size 50707856
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ae32f7a6c43d40740199b20c1720539f3d6dcd73d1bc0ff0c9d22f4386bdc37
3
+ size 50348480
onnx/decoder_with_past_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b0c3cfb8a019b7524a3d9c7ca8d47bc6a7ee3449c8413f105e398c1be11cbae0
3
- size 49458544
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac92773054aca908fca90d7e84a8c97a995526e9fe6292cf00c6e77ff7d223dd
3
+ size 49114568
onnx/encoder_model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:00c5456c76e0fefded6980221be42f67470feebf6f7b510ea3e645c42b6bbf71
3
- size 10182475
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca9d7bb2836193704b7e2435e3bbadbed985ac3a79ab7406b244b8865ab1a5c0
3
+ size 10113248
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
+ "Transpose",
8
+ "Add",
9
+ "Pow",
10
+ "Reshape",
11
+ "Div",
12
+ "Sqrt",
13
+ "Gather",
14
+ "Shape",
15
+ "Conv",
16
+ "Mul",
17
+ "ReduceMean",
18
+ "Unsqueeze",
19
+ "Sub",
20
+ "Concat",
21
+ "Constant",
22
+ "MatMul",
23
+ "Erf",
24
+ "Softmax"
25
+ ],
26
+ "weight_type": "QUInt8"
27
+ },
28
+ "decoder_model": {
29
+ "op_types": [
30
+ "Reshape",
31
+ "Equal",
32
+ "MatMul",
33
+ "Constant",
34
+ "Add",
35
+ "Pow",
36
+ "ReduceMean",
37
+ "Less",
38
+ "Cast",
39
+ "Concat",
40
+ "ConstantOfShape",
41
+ "Sqrt",
42
+ "Gather",
43
+ "Div",
44
+ "Sub",
45
+ "Where",
46
+ "Transpose",
47
+ "Shape",
48
+ "Range",
49
+ "Squeeze",
50
+ "Mul",
51
+ "Expand",
52
+ "Unsqueeze",
53
+ "Slice",
54
+ "Erf",
55
+ "Softmax"
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
+ "Transpose",
68
+ "Add",
69
+ "Reshape",
70
+ "Pow",
71
+ "Sqrt",
72
+ "Gather",
73
+ "Div",
74
+ "Shape",
75
+ "ReduceMean",
76
+ "Mul",
77
+ "Unsqueeze",
78
+ "Sub",
79
+ "MatMul",
80
+ "Constant",
81
+ "Concat",
82
+ "Slice",
83
+ "Erf",
84
+ "Softmax"
85
+ ],
86
+ "weight_type": "QInt8"
87
+ }
88
+ }
89
+ }