Skip to content

tmp ada #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
caizhi-mt wants to merge 1 commit into
base: mtgpu_hvd
Choose a base branch
from
Open

tmp ada #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -146,7 +146,7 @@
# -----------------------------------------------------------------------------
_C.TRAIN = CN()
_C.TRAIN.START_EPOCH = 0
_C.TRAIN.EPOCHS = 300
_C.TRAIN.EPOCHS = 90
_C.TRAIN.WARMUP_EPOCHS = 20
_C.TRAIN.WEIGHT_DECAY = 0.05
_C.TRAIN.BASE_LR = 5e-4
Expand Down
2 changes: 1 addition & 1 deletion models/swin_transformer_v2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -153,7 +153,7 @@ def forward(self, x, mask=None):

# cosine attention
attn = (F.normalize(q, dim=-1) @ F.normalize(k, dim=-1).transpose(-2, -1))
logit_scale = torch.clamp(self.logit_scale, max=torch.log(torch.tensor(1. / 0.01))).exp()
logit_scale = torch.clamp(self.logit_scale, max=torch.log(torch.tensor(1. / 0.01, device = "cuda"))).exp()
attn = attn * logit_scale

relative_position_bias_table = self.cpb_mlp(self.relative_coords_table).view(-1, self.num_heads)
Expand Down
2 changes: 1 addition & 1 deletion train.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1 +1 @@
horovodrun -np 4 python main_hvd.py --cfg configs/swinv2/swinv2_base_patch4_window8_128_hvd.yaml --data-path ../../datasets/tiny-imagenet-200 --batch-size 64 --disable_amp --local_rank 0
horovodrun -np 8 python main_hvd.py --cfg configs/swinv2/swinv2_base_patch4_window8_128_hvd.yaml --data-path ../tiny-imagenet-200 --batch-size 32 --disable_amp --local_rank 0