github/AMD-SHARK-Studio
1
1
Fork 0
mirror of https://github.com/nod-ai/AMD-SHARK-Studio.git synced 2026-04-03 03:00:17 -04:00
Code Issues Packages Projects Releases Wiki Activity

Add bert-large-uncased Torch model

Browse Source
This commit is contained in:
mariecwhite
2023-02-15 21:02:56 -08:00
parent 13a255afad
commit aac5de5b02
3 changed files with 3 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
tank/all_models.csv
View File

@@ -18,6 +18,7 @@ alexnet,linalg,torch,1e-2,1e-3,default,None,True,True,False,"https://github.com/
bert-base-cased,linalg,torch,1e-2,1e-3,default,None,False,False,False,"",""
bert-base-uncased,linalg,torch,1e-2,1e-3,default,None,False,False,False,"",""
bert-base-uncased_fp16,linalg,torch,1e-1,1e-1,default,None,True,False,True,"",""
bert-large-uncased,linalg,torch,1e-2,1e-3,default,None,False,False,False,"",""
facebook/deit-small-distilled-patch16-224,linalg,torch,1e-2,1e-3,default,nhcw-nhwc,False,True,False,"Fails during iree-compile.",""
google/vit-base-patch16-224,linalg,torch,1e-2,1e-3,default,nhcw-nhwc,False,True,False,"https://github.com/nod-ai/SHARK/issues/311",""
microsoft/beit-base-patch16-224-pt22k-ft22k,linalg,torch,1e-2,1e-3,default,nhcw-nhwc,False,True,False,"https://github.com/nod-ai/SHARK/issues/390",""
1 resnet50 mhlo tf 1e-2 1e-3 default nhcw-nhwc False False False macos
18 bert-base-cased linalg torch 1e-2 1e-3 default None False False False
19 bert-base-uncased linalg torch 1e-2 1e-3 default None False False False
20 bert-base-uncased_fp16 linalg torch 1e-1 1e-1 default None True False True
21 bert-large-uncased linalg torch 1e-2 1e-3 default None False False False
22 facebook/deit-small-distilled-patch16-224 linalg torch 1e-2 1e-3 default nhcw-nhwc False True False Fails during iree-compile.
23 google/vit-base-patch16-224 linalg torch 1e-2 1e-3 default nhcw-nhwc False True False https://github.com/nod-ai/SHARK/issues/311
24 microsoft/beit-base-patch16-224-pt22k-ft22k linalg torch 1e-2 1e-3 default nhcw-nhwc False True False https://github.com/nod-ai/SHARK/issues/390

1
tank/model_metadata.csv
View File

@@ -31,3 +31,4 @@ xlm-roberta-base,False,False,-,-,-
facebook/convnext-tiny-224,False,False,-,-,-
efficientnet-v2-s,False,False,22M,"image-classification,cnn","Includes MBConv and Fused-MBConv"
mnasnet1_0,False,True,-,"cnn, torchvision, mobile, architecture-search","Outperforms other mobile CNNs on Accuracy vs. Latency"
bert-large-uncased,True,hf,True,330M,"nlp;bert-variant;transformer-encoder","24 layers, 1024 hidden units, 16 attention heads"
1 model_name use_tracing dynamic param_count tags notes
31 facebook/convnext-tiny-224 False False - - -
32 efficientnet-v2-s False False 22M image-classification,cnn Includes MBConv and Fused-MBConv
33 mnasnet1_0 False True - cnn, torchvision, mobile, architecture-search Outperforms other mobile CNNs on Accuracy vs. Latency
34 bert-large-uncased True hf True 330M nlp;bert-variant;transformer-encoder

1
tank/torch_model_list.csv
View File

@@ -18,3 +18,4 @@ nvidia/mit-b0,True,hf_img_cls,False,3.7M,"image-classification,transformer-encod
mnasnet1_0,False,vision,True,-,"cnn, torchvision, mobile, architecture-search","Outperforms other mobile CNNs on Accuracy vs. Latency"
resnet50_fp16,False,vision,True,23M,"cnn,image-classification,residuals,resnet-variant","Bottlenecks with only conv2d (1x1 conv -> 3x3 conv -> 1x1 conv blocks)"
bert-base-uncased_fp16,True,fp16,False,109M,"nlp;bert-variant;transformer-encoder","12 layers; 768 hidden; 12 attention heads"
bert-large-uncased,True,hf,True,330M,"nlp;bert-variant;transformer-encoder","24 layers, 1024 hidden units, 16 attention heads"
1 model_name use_tracing model_type dynamic param_count tags notes
18 mnasnet1_0 False vision True - cnn, torchvision, mobile, architecture-search Outperforms other mobile CNNs on Accuracy vs. Latency
19 resnet50_fp16 False vision True 23M cnn,image-classification,residuals,resnet-variant Bottlenecks with only conv2d (1x1 conv -> 3x3 conv -> 1x1 conv blocks)
20 bert-base-uncased_fp16 True fp16 False 109M nlp;bert-variant;transformer-encoder 12 layers; 768 hidden; 12 attention heads
21 bert-large-uncased True hf True 330M nlp;bert-variant;transformer-encoder 24 layers, 1024 hidden units, 16 attention heads