Multiscale Vision Transformers #40

Amapocho · 2021-11-28T06:19:51Z

codecov-commenter · 2021-11-28T06:24:19Z

Codecov Report

Merging #40 (c2bffb8) into main (a8e4215) will decrease coverage by 11.15%.
The diff coverage is 0.00%.

@@             Coverage Diff              @@
##              main      #40       +/-   ##
============================================
- Coverage   100.00%   88.84%   -11.16%     
============================================
  Files           49       52        +3     
  Lines         1091     1228      +137     
============================================
  Hits          1091     1091               
- Misses           0      137      +137

Impacted Files	Coverage Δ
vformer/attention/multiscale.py	`0.00% <0.00%> (ø)`
vformer/encoder/embedding/patch_multiscale.py	`0.00% <0.00%> (ø)`
vformer/encoder/multiscale.py	`0.00% <0.00%> (ø)`

abhi-glitchhg

Could you please use Feedforward class from encoder/nn as MLP Instead of 'common/mlp'?

Also you can import DropPath from timm library, use this import statement - from timm.models.layers import DropPath,

Also if Pull Resquest is still in progress, you can change it to Draft.

Amapocho · 2021-11-28T07:14:30Z

Could you please use Feedforward class from encoder/nn as MLP Instead of 'common/mlp'?

Also you can import DropPath from timm library, use this import statement - from timm.models.layers import DropPath,

Also if Pull Resquest is still in progress, you can change it to Draft.

Made the necessary changes

vformer/encoder/embedding/patch_multiscale.py

abhi-glitchhg · 2022-01-14T19:24:24Z

vformer/encoder/multiscale.py

+
+    def forward(self, x, thw_shape):
+        x_block, thw_shape_new = self.attn(self.norm1(x), thw_shape)
+        x_res, _ = attention_pool(


here You need to import attention_pool from the attention module.

Co-authored-by: Abhijit Deo <[email protected]>

vformer/models/classification/multiscale.py

abhi-glitchhg · 2022-01-16T19:49:04Z

vformer/models/classification/multiscale.py

+                    POOL_KV_STRIDE[i][0]
+                ] = KVQ_KERNEL


KVQ_KERNEL this variable is not defined .

vformer/models/classification/multiscale.py

Co-authored-by: Abhijit Deo <[email protected]>

Amapocho added 13 commits November 27, 2021 01:11

Create multiscale.py

9b717ff

Update multiscale.py

f35e6e8

Create mlp.py

a3f6095

Delete mlp.py

441df76

Create droppath.py

03fc0f1

Create mlp.py

cd159f8

Add docstrings

7aef807

Update Mlp and DropPath import

5c8a00e

Update docstring

fb41ba3

Add import statements

9aa3800

Add docstring

6d38d9d

Merge branch 'SforAiDl:main' into main

38d9bc0

Run pre-commit

3788eb6

abhi-glitchhg requested changes Nov 28, 2021

View reviewed changes

Amapocho changed the title ~~Multiscale Vision Transformers~~ Multiscale Vision Transformers [Draft] Nov 28, 2021

Amapocho added 4 commits November 28, 2021 12:40

Update DropPath import

db284a8

Delete droppath.py

60aa249

Delete mlp.py

8df5268

Update Mlp import

5a08e33

Amapocho added 5 commits November 28, 2021 12:49

Run pre-commit

dd72553

Update docstring

f144ed7

Update mlp usage for feedforward

77e9925

Run pre-commit

d6fcc13

Update docstring

d716b63

Amapocho changed the title ~~Multiscale Vision Transformers [Draft]~~ Multiscale Vision Transformers Nov 28, 2021

Amapocho marked this pull request as draft November 28, 2021 08:50

Amapocho added 2 commits November 28, 2021 16:38

Remove multiscale block

f857764

Update imports

23a81f8

Amapocho added 5 commits January 14, 2022 14:08

Update test_encoder.py

0e61971

Update multiscale.py

313b2d0

Update multiscale.py

8a9a06c

Update test_models.py

36010fd

update

558dcff

abhi-glitchhg reviewed Jan 14, 2022

View reviewed changes

Amapocho and others added 3 commits January 15, 2022 01:03

Update vformer/encoder/embedding/patch_multiscale.py

e555aa0

Co-authored-by: Abhijit Deo <[email protected]>

Update tests

6fac3a7

Update tests

7cc4cea

abhi-glitchhg requested changes Jan 16, 2022

View reviewed changes

Amapocho and others added 19 commits January 17, 2022 07:56

Update vformer/models/classification/multiscale.py

077e2e7

Co-authored-by: Abhijit Deo <[email protected]>

Update vformer/models/classification/multiscale.py

6fea53d

Co-authored-by: Abhijit Deo <[email protected]>

Update vformer/models/classification/multiscale.py

e43a858

Co-authored-by: Abhijit Deo <[email protected]>

Update vformer/models/classification/multiscale.py

372aa69

Co-authored-by: Abhijit Deo <[email protected]>

Update vformer/models/classification/multiscale.py

5ad17bd

Co-authored-by: Abhijit Deo <[email protected]>

Update vformer/models/classification/multiscale.py

8b0b82e

Co-authored-by: Abhijit Deo <[email protected]>

Import trunc_normal_

a0a2e18

Create multiscale.py

14ddfb8

Update multiscale.py

5b603a2

Update multiscale.py

c40f844

Update __init__.py

9c14715

Update multiscale.py

3947d2c

Update multiscale.py

014c6e5

Update multiscale.py

85fd771

Update

b43eadb

Update multiscale.py

161ca3f

remove logger

beed6ea

Update multiscale.py

4926fed

Update HEAD_ACT

209b510

NeelayS closed this Jul 3, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Multiscale Vision Transformers #40

Multiscale Vision Transformers #40

Amapocho commented Nov 28, 2021

codecov-commenter commented Nov 28, 2021 •

edited

Loading

abhi-glitchhg left a comment

Amapocho commented Nov 28, 2021

abhi-glitchhg Jan 14, 2022 •

edited

Loading

abhi-glitchhg Jan 16, 2022

Multiscale Vision Transformers #40

Multiscale Vision Transformers #40

Conversation

Amapocho commented Nov 28, 2021

codecov-commenter commented Nov 28, 2021 • edited Loading

Codecov Report

abhi-glitchhg left a comment

Choose a reason for hiding this comment

Amapocho commented Nov 28, 2021

abhi-glitchhg Jan 14, 2022 • edited Loading

Choose a reason for hiding this comment

abhi-glitchhg Jan 16, 2022

Choose a reason for hiding this comment

codecov-commenter commented Nov 28, 2021 •

edited

Loading

abhi-glitchhg Jan 14, 2022 •

edited

Loading