@@ -13,28 +13,111 @@ across various devices by running comprehensive GitHub workflows.
1313
1414<!-- Torchbenchmark start -->
1515
16- | | Ascend NPU | Intel Gaudi |
17- | ------------------------| ------------| -------------|
18- | alexnet | ✅ | ✅ |
19- | BERT_pytorch | ✅ | ✅ |
20- | dcgan | ✅ | ✅ |
21- | hf_clip | ✅ | ✅ |
22- | hf_GPT2 | ✅ | ✅ |
23- | hf_GPT2_large | ✅ | ✅ |
24- | hf_Reformer | ⚠️ | ✅ |
25- | hf_Whisper | ✅ | ✅ |
26- | llama | ✅ | ✅ |
27- | llama_v2_7b_16h | ❌ | ❌ |
28- | llava | ⚠️ | ✅ |
29- | moondream | ❌ | ❌ |
30- | nanogpt | ✅ | ✅ |
31- | nvidia_deeprecommender | ❌ | ❌ |
32- | opacus_cifar10 | ✅ | ✅ |
33- | resnet152 | ✅ | ✅ |
34- | resnet18 | ✅ | ✅ |
35- | resnet50 | ✅ | ✅ |
36- | vgg16 | ✅ | ✅ |
37- | yolov3 | ✅ | ✅ |
16+ | | [ torch_npu] [ 1 ] |
17+ | ---------------------------------| ----------------|
18+ | simple_gpt | ❌ |
19+ | detectron2_fasterrcnn_r_50_dc5 | ❌ |
20+ | LearningToPaint | ✅ |
21+ | hf_GPT2_large | ✅ |
22+ | dcgan | ✅ |
23+ | nanogpt | ✅ |
24+ | fastNLP_Bert | ✅ |
25+ | moondream | ❌ |
26+ | mobilenet_v2_quantized_qat | ❌ |
27+ | functorch_dp_cifar10 | ✅ |
28+ | simple_gpt_tp_manual | ❌ |
29+ | speech_transformer | ✅ |
30+ | yolov3 | ✅ |
31+ | resnet50_quantized_qat | ❌ |
32+ | sam_fast | ❌ |
33+ | alexnet | ✅ |
34+ | timm_efficientnet | ✅ |
35+ | pyhpc_isoneutral_mixing | ✅ |
36+ | basic_gnn_edgecnn | ✅ |
37+ | nvidia_deeprecommender | ❌ |
38+ | opacus_cifar10 | ✅ |
39+ | dlrm | ✅ |
40+ | hf_Bert | ✅ |
41+ | hf_T5_generate | ✅ |
42+ | resnet50 | ✅ |
43+ | hf_BigBird | ✅ |
44+ | resnext50_32x4d | ✅ |
45+ | pyhpc_turbulent_kinetic_energy | ✅ |
46+ | llama | ✅ |
47+ | detectron2_maskrcnn_r_50_c4 | ❌ |
48+ | Super_SloMo | ✅ |
49+ | moco | ❌ |
50+ | stable_diffusion_unet | ❌ |
51+ | microbench_unbacked_tolist_sum | ✅ |
52+ | detectron2_maskrcnn_r_101_c4 | ❌ |
53+ | hf_distil_whisper | ✅ |
54+ | mnasnet1_0 | ✅ |
55+ | detectron2_fasterrcnn_r_50_fpn | ❌ |
56+ | timm_resnest | ✅ |
57+ | hf_GPT2 | ✅ |
58+ | squeezenet1_1 | ✅ |
59+ | basic_gnn_gin | ✅ |
60+ | hf_clip | ✅ |
61+ | mobilenet_v2 | ✅ |
62+ | drq | ✅ |
63+ | hf_Roberta_base | ✅ |
64+ | detectron2_maskrcnn_r_50_fpn | ❌ |
65+ | timm_nfnet | ✅ |
66+ | timm_vovnet | ✅ |
67+ | doctr_det_predictor | ✅ |
68+ | sam | ✅ |
69+ | hf_T5_large | ✅ |
70+ | mobilenet_v3_large | ✅ |
71+ | detectron2_fcos_r_50_fpn | ❌ |
72+ | soft_actor_critic | ✅ |
73+ | llava | ❌ |
74+ | timm_regnet | ✅ |
75+ | functorch_maml_omniglot | ✅ |
76+ | detectron2_fasterrcnn_r_101_c4 | ❌ |
77+ | hf_DistilBert | ✅ |
78+ | tts_angular | ✅ |
79+ | detectron2_maskrcnn | ❌ |
80+ | basic_gnn_sage | ✅ |
81+ | tacotron2 | ❌ |
82+ | detectron2_maskrcnn_r_101_fpn | ❌ |
83+ | lennard_jones | ✅ |
84+ | pytorch_unet | ✅ |
85+ | vgg16 | ✅ |
86+ | BERT_pytorch | ✅ |
87+ | timm_efficientdet | ❌ |
88+ | pyhpc_equation_of_state | ✅ |
89+ | maml | ✅ |
90+ | detectron2_fasterrcnn_r_50_c4 | ❌ |
91+ | resnet152 | ✅ |
92+ | phlippe_densenet | ✅ |
93+ | maml_omniglot | ✅ |
94+ | phlippe_resnet | ✅ |
95+ | pytorch_CycleGAN_and_pix2pix | ✅ |
96+ | hf_Whisper | ✅ |
97+ | hf_T5 | ✅ |
98+ | densenet121 | ✅ |
99+ | cm3leon_generate | ✅ |
100+ | detectron2_fasterrcnn_r_101_fpn | ❌ |
101+ | hf_Bert_large | ✅ |
102+ | stable_diffusion_text_encoder | ❌ |
103+ | hf_Reformer | ❌ |
104+ | detectron2_fasterrcnn_r_101_dc5 | ❌ |
105+ | demucs | ✅ |
106+ | pytorch_stargan | ✅ |
107+ | hf_T5_base | ✅ |
108+ | torch_multimodal_clip | ✅ |
109+ | vision_maskrcnn | ❌ |
110+ | timm_vision_transformer_large | ✅ |
111+ | hf_Bart | ✅ |
112+ | shufflenet_v2_x1_0 | ✅ |
113+ | llama_v2_7b_16h | ❌ |
114+ | basic_gnn_gcn | ✅ |
115+ | resnet18 | ✅ |
116+ | Background_Matting | ✅ |
117+ | doctr_reco_predictor | ✅ |
118+ | timm_vision_transformer | ✅ |
119+ | hf_Albert | ✅ |
120+ | hf_Longformer | ✅ |
38121
39122[ 1 ] : https://github.com/ascend/pytorch
40123
0 commit comments