[E2E_baseline] Dynamo Benchmark E2E Accuracy test timm_models has some fail_accuracy models

[chuanqi129]

🐛 Describe the bug

Timm_models has some models failed on accuracy check, the detail model list can be found as below table.

precision	mode	model
bfloat16	inference	eca_botnext26ts_256
bfloat16	inference	levit_128
bfloat16	inference	adv_inception_v3
bfloat16	inference	nfnet_l0
bfloat16	inference	eca_halonext26ts
bfloat16	inference	mixer_b16_224
bfloat16	inference	gmixer_24_224
bfloat16	inference	beit_base_patch16_224
bfloat16	inference	swin_base_patch4_window7_224
bfloat16	inference	crossvit_9_240
bfloat16	inference	ese_vovnet19b_dw
bfloat16	inference	gmlp_s16_224
bfloat16	inference	resmlp_12_224
bfloat16	inference	resnest101e
bfloat16	inference	mixnet_l
bfloat16	inference	swsl_resnext101_32x16d
bfloat16	inference	dm_nfnet_f0
bfloat16	inference	lcnet_050
bfloat16	inference	regnety_002
bfloat16	inference	tf_efficientnet_b0
bfloat16	inference	sebotnet33ts_256
bfloat16	inference	fbnetv3_b
bfloat16	inference	tf_mixnet_l
bfloat16	inference	mobilenetv3_large_100
bfloat16	inference	tinynet_a
bfloat16	inference	ghostnet_100
bfloat16	inference	twins_pcpvt_base
bfloat16	training	levit_128
bfloat16	training	eca_botnext26ts_256
bfloat16	training	gluon_inception_v3
bfloat16	training	adv_inception_v3
bfloat16	training	mixer_b16_224
bfloat16	training	nfnet_l0
bfloat16	training	res2net50_14w_8s
bfloat16	training	eca_halonext26ts
bfloat16	training	beit_base_patch16_224
bfloat16	training	gmixer_24_224
bfloat16	training	botnet26t_256
bfloat16	training	gmlp_s16_224
bfloat16	training	res2next50
bfloat16	training	crossvit_9_240
bfloat16	training	resmlp_12_224
bfloat16	training	swin_base_patch4_window7_224
bfloat16	training	hrnet_w18
bfloat16	training	cspdarknet53
bfloat16	training	inception_v3
bfloat16	training	resnest101e
bfloat16	training	dla102
bfloat16	training	dm_nfnet_f0
bfloat16	training	regnety_002
bfloat16	training	lcnet_050
bfloat16	training	repvgg_a2
bfloat16	training	sebotnet33ts_256
bfloat16	training	selecsls42b
bfloat16	training	res2net101_26w_4s
bfloat16	training	mobilenetv3_large_100
bfloat16	training	gernet_l
bfloat16	training	tf_mixnet_l
bfloat16	training	mobilevit_s
bfloat16	training	ghostnet_100
bfloat16	training	tinynet_a
bfloat16	training	twins_pcpvt_base
bfloat16	training	visformer_small
bfloat16	training	volo_d1_224
float16	inference	eca_botnext26ts_256
float16	inference	levit_128
float16	inference	res2net50_14w_8s
float16	inference	nfnet_l0
float16	inference	gmixer_24_224
float16	inference	eca_halonext26ts
float16	inference	mixer_b16_224
float16	inference	beit_base_patch16_224
float16	inference	res2next50
float16	inference	swin_base_patch4_window7_224
float16	inference	ese_vovnet19b_dw
float16	inference	crossvit_9_240
float16	inference	gmlp_s16_224
float16	inference	resmlp_12_224
float16	inference	resnest101e
float16	inference	mixnet_l
float16	inference	inception_v3
float16	inference	dm_nfnet_f0
float16	inference	lcnet_050
float16	inference	poolformer_m36
float16	inference	regnety_002
float16	inference	sebotnet33ts_256
float16	inference	tf_efficientnet_b0
float16	inference	fbnetv3_b
float16	inference	mobilenetv3_large_100
float16	inference	tf_mixnet_l
float16	inference	ghostnet_100
float16	inference	tinynet_a
float16	inference	twins_pcpvt_base
float16	training	levit_128
float16	training	eca_botnext26ts_256
float16	training	gluon_inception_v3
float16	training	adv_inception_v3
float16	training	mixer_b16_224
float16	training	nfnet_l0
float16	training	res2net50_14w_8s
float16	training	eca_halonext26ts
float16	training	gmixer_24_224
float16	training	beit_base_patch16_224
float16	training	gmlp_s16_224
float16	training	botnet26t_256
float16	training	res2next50
float16	training	ese_vovnet19b_dw
float16	training	crossvit_9_240
float16	training	resmlp_12_224
float16	training	swin_base_patch4_window7_224
float16	training	hrnet_w18
float16	training	cspdarknet53
float16	training	resnest101e
float16	training	inception_v3
float16	training	dla102
float16	training	poolformer_m36
float16	training	dm_nfnet_f0
float16	training	regnety_002
float16	training	lcnet_050
float16	training	sebotnet33ts_256
float16	training	repvgg_a2
float16	training	selecsls42b
float16	training	res2net101_26w_4s
float16	training	mobilenetv3_large_100
float16	training	gernet_l
float16	training	tf_mixnet_l
float16	training	mobilevit_s
float16	training	ghostnet_100
float16	training	tinynet_a
float16	training	twins_pcpvt_base
float16	training	visformer_small
float16	training	volo_d1_224
float32	training	levit_128
float32	training	eca_botnext26ts_256
float32	training	gluon_inception_v3
float32	training	adv_inception_v3
float32	training	res2net50_14w_8s
float32	training	eca_halonext26ts
float32	training	botnet26t_256
float32	training	res2next50
float32	training	ese_vovnet19b_dw
float32	training	cspdarknet53
float32	training	hrnet_w18
float32	training	resnest101e
float32	training	inception_v3
float32	training	dla102
float32	training	regnety_002
float32	training	repvgg_a2
float32	training	lcnet_050
float32	training	sebotnet33ts_256
float32	training	selecsls42b
float32	training	res2net101_26w_4s
float32	training	mobilenetv3_large_100
float32	training	gernet_l
float32	training	tf_mixnet_l
float32	training	mobilevit_s
float32	training	ghostnet_100
float32	training	tinynet_a
float32	training	visformer_small
float32	training	volo_d1_224

Versions

Pytorch: git clone -b e2e-baseline https://github.com/etaf/pytorch-inductor-xpu pytorch Test script: inductor_xpu_test.sh

[riverliuintel]

@etaf have a look whether there is common issue. Please have the first round of check. For details model failure triage, Yunfei will cover it.