Skip to content

Commit ba62990

Browse files
authored
Update output for quantized models by ORT on VNNI machine (#572)
* detect vnni Signed-off-by: jcwchen <[email protected]> * only consider quant models Signed-off-by: jcwchen <[email protected]> * model_path Signed-off-by: jcwchen <[email protected]> * update output with VNNI machine Signed-off-by: Chun-Wei Chen <[email protected]> * at least run inferencesession Signed-off-by: jcwchen <[email protected]> * update manifest Signed-off-by: jcwchen <[email protected]> Signed-off-by: jcwchen <[email protected]> Signed-off-by: Chun-Wei Chen <[email protected]>
1 parent 98fd2a6 commit ba62990

File tree

16 files changed

+59
-81
lines changed

16 files changed

+59
-81
lines changed

ONNX_HUB_MANIFEST.json

Lines changed: 26 additions & 50 deletions
Original file line numberDiff line numberDiff line change
@@ -151,8 +151,8 @@
151151
]
152152
},
153153
"model_with_data_path": "text/machine_comprehension/bert-squad/model/bertsquad-12-int8.tar.gz",
154-
"model_with_data_sha": "5f4bd418f2ee55310788fed78842297ec0b3ecda6669c563196f20396cb4d401",
155-
"model_with_data_bytes": 105881973
154+
"model_with_data_sha": "6c33f44ee1949ee25936e259cb68da3e19bea7f6e7d5ea72a4d95ae300c86d87",
155+
"model_with_data_bytes": 106044512
156156
}
157157
},
158158
{
@@ -2041,8 +2041,8 @@
20412041
]
20422042
},
20432043
"model_with_data_path": "vision/classification/efficientnet-lite4/model/efficientnet-lite4-11-int8.tar.gz",
2044-
"model_with_data_sha": "3b03546fb343ba48e3243c893ea37e47cb1f337d80ed0aca562077d8d3da51b6",
2045-
"model_with_data_bytes": 12780764
2044+
"model_with_data_sha": "d3f50f4562e61d870a70e68157ccf64d8f03f6927540ad343e6fbddb79647e1f",
2045+
"model_with_data_bytes": 12789970
20462046
}
20472047
},
20482048
{
@@ -2083,8 +2083,8 @@
20832083
]
20842084
},
20852085
"model_with_data_path": "vision/classification/efficientnet-lite4/model/efficientnet-lite4-11-qdq.tar.gz",
2086-
"model_with_data_sha": "f58754ff53120d937e4bc32e6a815fc0163261f7b842762868ed453947258390",
2087-
"model_with_data_bytes": 10193490
2086+
"model_with_data_sha": "4802f078c57e328004e326623c3de9e0e9046d9ce485eb79ed05d4d3c478b417",
2087+
"model_with_data_bytes": 10194220
20882088
}
20892089
},
20902090
{
@@ -2560,8 +2560,8 @@
25602560
]
25612561
},
25622562
"model_with_data_path": "vision/classification/inception_and_googlenet/inception_v1/model/inception-v1-12-int8.tar.gz",
2563-
"model_with_data_sha": "296e49e0f1253877053d5aa48f4192f675b302b5828e70a821d6d94e0c8ee4e3",
2564-
"model_with_data_bytes": 9463738
2563+
"model_with_data_sha": "bc3d78134d9777d3e3a3320523834beb2ba7bdd1c5e017cba5a8d268691afbc6",
2564+
"model_with_data_bytes": 9474526
25652565
}
25662566
},
25672567
{
@@ -3311,8 +3311,8 @@
33113311
]
33123312
},
33133313
"model_with_data_path": "vision/classification/mobilenet/model/mobilenetv2-12-int8.tar.gz",
3314-
"model_with_data_sha": "f00ea406993d4b2a404a4c5f4258e862f68aedfbcc324510267e556612501211",
3315-
"model_with_data_bytes": 3910933
3314+
"model_with_data_sha": "dfcf1dc15b78611489f22a19902de12e304bf0380e185bde499fcf70f7647455",
3315+
"model_with_data_bytes": 3914892
33163316
}
33173317
},
33183318
{
@@ -3353,8 +3353,8 @@
33533353
]
33543354
},
33553355
"model_with_data_path": "vision/classification/mobilenet/model/mobilenetv2-12-qdq.tar.gz",
3356-
"model_with_data_sha": "29be79b1561b1bf028a064eb4061b23b92883cb7f22b50f5f52d685efd23a04c",
3357-
"model_with_data_bytes": 3431250
3356+
"model_with_data_sha": "dd8cafda6fdf773a37ba09858482232eca8e01997ec6f1bebf1aca11aea0c1ce",
3357+
"model_with_data_bytes": 3434401
33583358
}
33593359
},
33603360
{
@@ -4278,7 +4278,7 @@
42784278
}
42794279
},
42804280
{
4281-
"model": "ResNet50_int8",
4281+
"model": "ResNet50-int8",
42824282
"model_path": "vision/classification/resnet/model/resnet50-v1-12-int8.onnx",
42834283
"onnx_version": "1.7.0",
42844284
"opset_version": 12,
@@ -4314,20 +4314,9 @@
43144314
}
43154315
]
43164316
},
4317-
"extra_ports": {
4318-
"features": [
4319-
{
4320-
"name": "flatten_473_quantized",
4321-
"shape": [
4322-
0,
4323-
2048
4324-
]
4325-
}
4326-
]
4327-
},
43284317
"model_with_data_path": "vision/classification/resnet/model/resnet50-v1-12-int8.tar.gz",
4329-
"model_with_data_sha": "7ddcf6ac49d0cc1ab0d3b7ecb6b81e295714c3e3de7f582fee85a7bebd8426e9",
4330-
"model_with_data_bytes": 22318352
4318+
"model_with_data_sha": "662710f50cafe29f44727b9cd8be5430fb06da18d3004cf5f9d9645c0dc39d71",
4319+
"model_with_data_bytes": 22355322
43314320
}
43324321
},
43334322
{
@@ -4368,12 +4357,12 @@
43684357
]
43694358
},
43704359
"model_with_data_path": "vision/classification/resnet/model/resnet50-v1-12-qdq.tar.gz",
4371-
"model_with_data_sha": "986160ffdcf88dedba54ee1fd8ea9597a91b11ccf16016f495bffe22e83ea40e",
4372-
"model_with_data_bytes": 17043120
4360+
"model_with_data_sha": "6a93ba624ecab599d3f8511b1409f5103beb2f1708a85955397a3132ca423d99",
4361+
"model_with_data_bytes": 17049336
43734362
}
43744363
},
43754364
{
4376-
"model": "ResNet50_fp32",
4365+
"model": "ResNet50-fp32",
43774366
"model_path": "vision/classification/resnet/model/resnet50-v1-12.onnx",
43784367
"onnx_version": "1.7.0",
43794368
"opset_version": 12,
@@ -4409,19 +4398,6 @@
44094398
}
44104399
]
44114400
},
4412-
"extra_ports": {
4413-
"features": [
4414-
{
4415-
"name": "resnetv17_pool1_fwd",
4416-
"shape": [
4417-
0,
4418-
2048,
4419-
1,
4420-
1
4421-
]
4422-
}
4423-
]
4424-
},
44254401
"model_with_data_path": "vision/classification/resnet/model/resnet50-v1-12.tar.gz",
44264402
"model_with_data_sha": "9391137cfc8fbec372d7a1e59e272d67550dab72d93cf7c7d6256782262599ea",
44274403
"model_with_data_bytes": 96559469
@@ -4840,8 +4816,8 @@
48404816
]
48414817
},
48424818
"model_with_data_path": "vision/classification/shufflenet/model/shufflenet-v2-12-int8.tar.gz",
4843-
"model_with_data_sha": "c746144bf4383cf0a3f1e3a4093eb46da625f0ce4db1e5657d6f8fd3f1cf9d50",
4844-
"model_with_data_bytes": 2485799
4819+
"model_with_data_sha": "f538889aa028f6fcc4db1dce9b0bf2e7e347a81d5fc28087d96758dc0131e8c7",
4820+
"model_with_data_bytes": 2488137
48454821
}
48464822
},
48474823
{
@@ -5348,8 +5324,8 @@
53485324
]
53495325
},
53505326
"model_with_data_path": "vision/classification/vgg/model/vgg16-12-int8.tar.gz",
5351-
"model_with_data_sha": "477a77f7ba31bc6bebe6e9824cb0108173d3bb0c54506d8c6663ea36eee7dfb4",
5352-
"model_with_data_bytes": 105964121
5327+
"model_with_data_sha": "75c08988b598e24173a21a7dc0b396ca9dc49afb63be35e5c993ac56b297e4cf",
5328+
"model_with_data_bytes": 106112057
53535329
}
53545330
},
53555331
{
@@ -6326,8 +6302,8 @@
63266302
]
63276303
},
63286304
"model_with_data_path": "vision/object_detection_segmentation/faster-rcnn/model/FasterRCNN-12-int8.tar.gz",
6329-
"model_with_data_sha": "599fcd738eb08d8d5daee9edb332f785c59f9a3c704292ba420a41c1b43404e2",
6330-
"model_with_data_bytes": 37955516
6305+
"model_with_data_sha": "d497a293e8486ec0a56f5276eb246b93280b7f31dc4e23a33db444d262384591",
6306+
"model_with_data_bytes": 38019008
63316307
}
63326308
},
63336309
{
@@ -7242,8 +7218,8 @@
72427218
]
72437219
},
72447220
"model_with_data_path": "vision/object_detection_segmentation/ssd/model/ssd-12-int8.tar.gz",
7245-
"model_with_data_sha": "9f70dc76cac7618375cdc2f4d4b8a42bea229d2545882df57a3b6cce282515a5",
7246-
"model_with_data_bytes": 31815039
7221+
"model_with_data_sha": "b8984dd954eb1463a8714d9f25d87d1d7276479a78660ce5ec37196a81d31413",
7222+
"model_with_data_bytes": 31845320
72477223
}
72487224
},
72497225
{
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:5f4bd418f2ee55310788fed78842297ec0b3ecda6669c563196f20396cb4d401
3-
size 105881973
2+
oid sha256:6c33f44ee1949ee25936e259cb68da3e19bea7f6e7d5ea72a4d95ae300c86d87
3+
size 106044512

text/machine_comprehension/gpt2-bs/model/gpt2-lm-head-bs-12.onnx

100755100644
File mode changed.
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:3b03546fb343ba48e3243c893ea37e47cb1f337d80ed0aca562077d8d3da51b6
3-
size 12780764
2+
oid sha256:d3f50f4562e61d870a70e68157ccf64d8f03f6927540ad343e6fbddb79647e1f
3+
size 12789970
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:f58754ff53120d937e4bc32e6a815fc0163261f7b842762868ed453947258390
3-
size 10193490
2+
oid sha256:4802f078c57e328004e326623c3de9e0e9046d9ce485eb79ed05d4d3c478b417
3+
size 10194220
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:296e49e0f1253877053d5aa48f4192f675b302b5828e70a821d6d94e0c8ee4e3
3-
size 9463738
2+
oid sha256:bc3d78134d9777d3e3a3320523834beb2ba7bdd1c5e017cba5a8d268691afbc6
3+
size 9474526
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:f00ea406993d4b2a404a4c5f4258e862f68aedfbcc324510267e556612501211
3-
size 3910933
2+
oid sha256:dfcf1dc15b78611489f22a19902de12e304bf0380e185bde499fcf70f7647455
3+
size 3914892
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:29be79b1561b1bf028a064eb4061b23b92883cb7f22b50f5f52d685efd23a04c
3-
size 3431250
2+
oid sha256:dd8cafda6fdf773a37ba09858482232eca8e01997ec6f1bebf1aca11aea0c1ce
3+
size 3434401
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:7ddcf6ac49d0cc1ab0d3b7ecb6b81e295714c3e3de7f582fee85a7bebd8426e9
3-
size 22318352
2+
oid sha256:662710f50cafe29f44727b9cd8be5430fb06da18d3004cf5f9d9645c0dc39d71
3+
size 22355322
Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,3 @@
11
version https://git-lfs.github.com/spec/v1
2-
oid sha256:986160ffdcf88dedba54ee1fd8ea9597a91b11ccf16016f495bffe22e83ea40e
3-
size 17043120
2+
oid sha256:6a93ba624ecab599d3f8511b1409f5103beb2f1708a85955397a3132ca423d99
3+
size 17049336

0 commit comments

Comments
 (0)