tarekziade
commited on
Commit
•
9b9dca5
1
Parent(s):
55c1ce0
Upload 19 files
Browse files- model.safetensors +1 -1
- onnx/decoder_model.onnx +2 -2
- onnx/decoder_model_merged.onnx +3 -0
- onnx/decoder_model_merged_quantized.onnx +3 -0
- onnx/decoder_model_quantized.onnx +2 -2
- onnx/decoder_with_past_model.onnx +3 -0
- onnx/decoder_with_past_model_quantized.onnx +3 -0
- onnx/encoder_model.onnx +1 -1
- onnx/encoder_model_quantized.onnx +2 -2
- quantize_config.json +93 -32
- training_args.bin +1 -1
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 729979160
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:dfbf6a06c87f2ecf5debe4ca7a967db958ef5ef91b9e95d0968db292b29bd044
|
3 |
size 729979160
|
onnx/decoder_model.onnx
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d7025996166a391f1a16c5087f33e58b13742635f068dcfba863f481229d3aa9
|
3 |
+
size 385863272
|
onnx/decoder_model_merged.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:67601b107ac8e081023020798926374cc1891e9b20888049b25af8b26e525036
|
3 |
+
size 387339536
|
onnx/decoder_model_merged_quantized.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fc5f9d3406b3311eea01659346b322c08c2ca993d6c4c9a5efef56894b0cb83e
|
3 |
+
size 99755449
|
onnx/decoder_model_quantized.onnx
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:365eb53fde6f242e34beec14cb7b1eeb9c80538fa1527cbe3104fcbaa5bde638
|
3 |
+
size 98063697
|
onnx/decoder_with_past_model.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f8b578768f4e7f43a54c96353cd8b06aed97c8d94411522906f021ca9195132b
|
3 |
+
size 385862852
|
onnx/decoder_with_past_model_quantized.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:89bab295c36f14f4dae4d6f3440f06c8620ca3b3aa671fa9e9c7114637eaa297
|
3 |
+
size 98061104
|
onnx/encoder_model.onnx
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 343440610
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fef6f2aa561891b1d39b2e79411e63cf7859ba2fa5a1a56376a13f7571c8f87a
|
3 |
size 343440610
|
onnx/encoder_model_quantized.onnx
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ea7ae423fd73637b434c8a3266ea049c25b09e321cef8d0c97138ba16ab01fda
|
3 |
+
size 87038151
|
quantize_config.json
CHANGED
@@ -2,63 +2,124 @@
|
|
2 |
"per_channel": false,
|
3 |
"reduce_range": false,
|
4 |
"per_model_config": {
|
5 |
-
"
|
6 |
"op_types": [
|
7 |
-
"
|
8 |
-
"
|
9 |
-
"Transpose",
|
10 |
"Reshape",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
"Where",
|
12 |
-
"
|
|
|
13 |
"ReduceMean",
|
14 |
-
"Pow",
|
15 |
-
"Constant",
|
16 |
"Tanh",
|
17 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
18 |
"Concat",
|
19 |
-
"
|
20 |
-
"
|
21 |
-
"
|
|
|
|
|
22 |
"Range",
|
23 |
-
"Gather",
|
24 |
"Shape",
|
25 |
-
"ConstantOfShape",
|
26 |
"Squeeze",
|
27 |
"Div",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
28 |
"Unsqueeze",
|
29 |
-
"
|
30 |
-
"
|
31 |
-
"
|
|
|
32 |
],
|
33 |
"weight_type": "QInt8"
|
34 |
},
|
35 |
"encoder_model": {
|
36 |
"op_types": [
|
37 |
-
"Slice",
|
38 |
-
"Transpose",
|
39 |
-
"Reshape",
|
40 |
-
"Where",
|
41 |
-
"Sqrt",
|
42 |
-
"ReduceMean",
|
43 |
-
"Pow",
|
44 |
"Constant",
|
45 |
-
"
|
|
|
|
|
46 |
"Concat",
|
47 |
-
"
|
48 |
-
"
|
49 |
-
"
|
50 |
"Equal",
|
|
|
|
|
|
|
51 |
"Gather",
|
|
|
|
|
52 |
"Conv",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
53 |
"Shape",
|
54 |
-
"
|
55 |
"Div",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
"Unsqueeze",
|
57 |
-
"
|
58 |
-
"
|
59 |
-
"
|
|
|
60 |
],
|
61 |
-
"weight_type": "
|
62 |
}
|
63 |
}
|
64 |
}
|
|
|
2 |
"per_channel": false,
|
3 |
"reduce_range": false,
|
4 |
"per_model_config": {
|
5 |
+
"decoder_with_past_model": {
|
6 |
"op_types": [
|
7 |
+
"Constant",
|
8 |
+
"Sqrt",
|
|
|
9 |
"Reshape",
|
10 |
+
"Concat",
|
11 |
+
"Cast",
|
12 |
+
"Split",
|
13 |
+
"Transpose",
|
14 |
+
"Add",
|
15 |
+
"Slice",
|
16 |
+
"Range",
|
17 |
+
"Shape",
|
18 |
+
"Squeeze",
|
19 |
+
"Div",
|
20 |
+
"ConstantOfShape",
|
21 |
+
"Gather",
|
22 |
+
"Pow",
|
23 |
+
"MatMul",
|
24 |
+
"Gemm",
|
25 |
+
"Softmax",
|
26 |
"Where",
|
27 |
+
"Unsqueeze",
|
28 |
+
"Mul",
|
29 |
"ReduceMean",
|
|
|
|
|
30 |
"Tanh",
|
31 |
+
"Sub"
|
32 |
+
],
|
33 |
+
"weight_type": "QInt8"
|
34 |
+
},
|
35 |
+
"decoder_model": {
|
36 |
+
"op_types": [
|
37 |
+
"Constant",
|
38 |
+
"Sqrt",
|
39 |
+
"Reshape",
|
40 |
"Concat",
|
41 |
+
"Cast",
|
42 |
+
"Split",
|
43 |
+
"Transpose",
|
44 |
+
"Add",
|
45 |
+
"Slice",
|
46 |
"Range",
|
|
|
47 |
"Shape",
|
|
|
48 |
"Squeeze",
|
49 |
"Div",
|
50 |
+
"ConstantOfShape",
|
51 |
+
"Gather",
|
52 |
+
"Pow",
|
53 |
+
"MatMul",
|
54 |
+
"Gemm",
|
55 |
+
"Softmax",
|
56 |
+
"Where",
|
57 |
"Unsqueeze",
|
58 |
+
"Mul",
|
59 |
+
"ReduceMean",
|
60 |
+
"Tanh",
|
61 |
+
"Sub"
|
62 |
],
|
63 |
"weight_type": "QInt8"
|
64 |
},
|
65 |
"encoder_model": {
|
66 |
"op_types": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
"Constant",
|
68 |
+
"Sqrt",
|
69 |
+
"Reshape",
|
70 |
+
"Erf",
|
71 |
"Concat",
|
72 |
+
"Transpose",
|
73 |
+
"Add",
|
74 |
+
"Slice",
|
75 |
"Equal",
|
76 |
+
"Shape",
|
77 |
+
"Div",
|
78 |
+
"ConstantOfShape",
|
79 |
"Gather",
|
80 |
+
"Pow",
|
81 |
+
"MatMul",
|
82 |
"Conv",
|
83 |
+
"Softmax",
|
84 |
+
"Where",
|
85 |
+
"Unsqueeze",
|
86 |
+
"Expand",
|
87 |
+
"Mul",
|
88 |
+
"ReduceMean",
|
89 |
+
"Sub"
|
90 |
+
],
|
91 |
+
"weight_type": "QUInt8"
|
92 |
+
},
|
93 |
+
"decoder_model_merged": {
|
94 |
+
"op_types": [
|
95 |
+
"Constant",
|
96 |
+
"Sqrt",
|
97 |
+
"Reshape",
|
98 |
+
"Concat",
|
99 |
+
"Cast",
|
100 |
+
"Split",
|
101 |
+
"Transpose",
|
102 |
+
"Add",
|
103 |
+
"Slice",
|
104 |
+
"Range",
|
105 |
"Shape",
|
106 |
+
"Squeeze",
|
107 |
"Div",
|
108 |
+
"ConstantOfShape",
|
109 |
+
"Gather",
|
110 |
+
"Pow",
|
111 |
+
"MatMul",
|
112 |
+
"If",
|
113 |
+
"Gemm",
|
114 |
+
"Softmax",
|
115 |
+
"Where",
|
116 |
"Unsqueeze",
|
117 |
+
"Mul",
|
118 |
+
"ReduceMean",
|
119 |
+
"Tanh",
|
120 |
+
"Sub"
|
121 |
],
|
122 |
+
"weight_type": "QInt8"
|
123 |
}
|
124 |
}
|
125 |
}
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5112
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fd3daca9a2e8e9dd775d2cd9f9a0e33ee454cf4e9d1bb12d250711d49ddadcfd
|
3 |
size 5112
|