ZF/vit-pytorch
1
0
Fork 0
mirror of https://github.com/lucidrains/vit-pytorch.git synced 2025-12-30 08:02:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

build @jrounds suggestion

Browse Source
This commit is contained in:
Phil Wang
2022-01-03 12:56:18 -08:00
parent 91ed738731
commit 0082301f9e
1 changed files with 8 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
vit_pytorch/mae.py
View File

@@ -14,11 +14,13 @@ class MAE(nn.Module):
masking_ratio = 0.75,
decoder_depth = 1,
decoder_heads = 8,
decoder_dim_head = 64
decoder_dim_head = 64,
apply_decoder_pos_emb_all = False # whether to (re)apply decoder positional embedding to encoder unmasked tokens
):
super().__init__()
assert masking_ratio > 0 and masking_ratio < 1, 'masking ratio must be kept between 0 and 1'
self.masking_ratio = masking_ratio
self.apply_decoder_pos_emb_all = apply_decoder_pos_emb_all
# extract some hyperparameters and functions from encoder (vision transformer to be trained)
@@ -71,6 +73,11 @@ class MAE(nn.Module):
decoder_tokens = self.enc_to_dec(encoded_tokens)
# reapply decoder position embedding to unmasked tokens, if desired
if self.apply_decoder_pos_emb_all:
decoder_tokens = decoder_tokens + self.decoder_pos_emb(unmasked_indices)
# repeat mask tokens for number of masked, and add the positions using the masked indices derived above
mask_tokens = repeat(self.mask_token, 'd -> b n d', b = batch, n = num_masked)