Aamir Xenova HF staff commited on
Commit
4afe992
1 Parent(s): 9b095c0

Add ONNX weights (#1)

Browse files

- Add ONNX weights (1a5e9052cb16a2a69e1b51aec90d2f335ef486b8)


Co-authored-by: Joshua <Xenova@users.noreply.huggingface.co>

onnx/model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4ab4c480f402e70429d67dcc80a2bd88476c166824dcb1d10602566c485fd0f
3
+ size 1741981193
onnx/model_quantized.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d03bd791409057b78184fa7726effdf5429114a68a7757939aa0d350761f7697
3
+ size 642585535
onnx/quantize_config.json ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_channel": true,
3
+ "reduce_range": true,
4
+ "per_model_config": {
5
+ "model": {
6
+ "op_types": [
7
+ "Ceil",
8
+ "Transpose",
9
+ "Abs",
10
+ "GatherElements",
11
+ "Sign",
12
+ "Unsqueeze",
13
+ "Pow",
14
+ "Log",
15
+ "Erf",
16
+ "Shape",
17
+ "Softmax",
18
+ "Div",
19
+ "Squeeze",
20
+ "Expand",
21
+ "Constant",
22
+ "Sqrt",
23
+ "Gemm",
24
+ "Less",
25
+ "ConstantOfShape",
26
+ "Range",
27
+ "Clip",
28
+ "LessOrEqual",
29
+ "MatMul",
30
+ "ReduceMean",
31
+ "Identity",
32
+ "Where",
33
+ "Gather",
34
+ "Greater",
35
+ "Neg",
36
+ "Equal",
37
+ "Concat",
38
+ "Add",
39
+ "Slice",
40
+ "Mul",
41
+ "Tile",
42
+ "Cast",
43
+ "Reshape",
44
+ "And",
45
+ "Sub"
46
+ ],
47
+ "weight_type": "QInt8"
48
+ }
49
+ }
50
+ }