Xenova HF staff commited on
Commit
c4f949b
1 Parent(s): 48e3b5c

Upload folder using huggingface_hub

Browse files
onnx/model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:19fb269e5df4a379a5486cb7682b9f348818fe7a81a6de1f9b2595ca75847dff
3
- size 813162230
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c39b27cfce3c3076be799058e8b9a7c53b77991cb5658ad29caa30d766e3e103
3
+ size 813154118
onnx/model_quantized.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1be8ca418e1a0792c4182a9f1a8a294de559cba61d258a55a3793bdb9ffa1904
3
- size 210988144
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9171eb00c38b9ec82f924877356d008b79e3285dbac7cd10965827bee30c9a99
3
+ size 210977441
quantize_config.json CHANGED
@@ -40,57 +40,57 @@
40
  },
41
  "text_model": {
42
  "op_types": [
43
- "Sub",
44
- "Div",
45
- "Gemm",
46
- "Add",
47
- "Constant",
48
  "Transpose",
49
- "Gather",
50
- "Tanh",
51
- "Mul",
52
  "MatMul",
53
- "Reshape",
54
  "Unsqueeze",
55
- "Softmax",
 
 
 
 
 
56
  "Shape",
57
- "Slice",
 
 
58
  "Concat",
59
- "Sqrt",
60
- "ReduceMean",
61
  "Cast",
62
- "Pow"
 
63
  ],
64
  "weight_type": "QInt8"
65
  },
66
  "vision_model": {
67
  "op_types": [
68
- "Sub",
 
69
  "ConstantOfShape",
 
 
70
  "Div",
71
- "Gemm",
72
  "Gather",
73
- "Add",
74
- "Transpose",
75
- "Constant",
76
- "Squeeze",
77
- "Tanh",
78
- "Mul",
79
- "Tile",
80
- "Conv",
81
- "MatMul",
82
  "Reshape",
83
- "Unsqueeze",
84
  "Expand",
85
- "Mod",
86
- "Softmax",
87
- "Slice",
88
  "Shape",
 
 
 
 
89
  "Concat",
90
- "Sqrt",
91
- "ReduceMean",
 
92
  "Cast",
93
- "Pow"
 
 
94
  ],
95
  "weight_type": "QUInt8"
96
  }
 
40
  },
41
  "text_model": {
42
  "op_types": [
 
 
 
 
 
43
  "Transpose",
 
 
 
44
  "MatMul",
45
+ "Slice",
46
  "Unsqueeze",
47
+ "Div",
48
+ "Sub",
49
+ "Gather",
50
+ "Reshape",
51
+ "Pow",
52
+ "Sqrt",
53
  "Shape",
54
+ "Mul",
55
+ "Tanh",
56
+ "Gemm",
57
  "Concat",
58
+ "Softmax",
59
+ "Constant",
60
  "Cast",
61
+ "Add",
62
+ "ReduceMean"
63
  ],
64
  "weight_type": "QInt8"
65
  },
66
  "vision_model": {
67
  "op_types": [
68
+ "Transpose",
69
+ "MatMul",
70
  "ConstantOfShape",
71
+ "Slice",
72
+ "Unsqueeze",
73
  "Div",
74
+ "Sub",
75
  "Gather",
 
 
 
 
 
 
 
 
 
76
  "Reshape",
77
+ "Squeeze",
78
  "Expand",
79
+ "Pow",
80
+ "Sqrt",
 
81
  "Shape",
82
+ "Mul",
83
+ "Tile",
84
+ "Tanh",
85
+ "Gemm",
86
  "Concat",
87
+ "Softmax",
88
+ "Mod",
89
+ "Constant",
90
  "Cast",
91
+ "Add",
92
+ "ReduceMean",
93
+ "Conv"
94
  ],
95
  "weight_type": "QUInt8"
96
  }