Skip to content
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

AttributeError: 'tuple' object has no attribute 'size' #195

Closed
wants to merge 4 commits into from
Closed

AttributeError: 'tuple' object has no attribute 'size' #195

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
87e3e11
fixes #141
Nov 10, 2022
c630a7a
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Nov 10, 2022
a10ba31
added missing transformer
Nov 10, 2022
1ddb8dc
Merge branch 'main' of github.com:fabiofumarola/torchinfo
Nov 10, 2022
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
1 change: 1 addition & 0 deletions requirements-dev.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,3 +8,4 @@ pylint
pytest
pytest-cov
pre-commit
transformers
46 changes: 46 additions & 0 deletions tests/test_output/huggingface.out
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,46 @@
==============================================================================================================
Layer (type:depth-idx) Output Shape Param #
==============================================================================================================
T5ForConditionalGeneration [2, 100, 512] --
├─T5Stack: 1-1 [2, 100, 512] 35,332,800
├─T5Stack: 1-2 -- (recursive)
│ └─Embedding: 2-1 [2, 100, 512] 16,449,536
├─T5Stack: 1-3 -- (recursive)
│ └─Dropout: 2-2 [2, 100, 512] --
│ └─ModuleList: 2-3 -- --
│ │ └─T5Block: 3-1 [2, 100, 512] 2,360,512
│ │ └─T5Block: 3-2 [2, 100, 512] 2,360,320
│ │ └─T5Block: 3-3 [2, 100, 512] 2,360,320
│ │ └─T5Block: 3-4 [2, 100, 512] 2,360,320
│ │ └─T5Block: 3-5 [2, 100, 512] 2,360,320
│ │ └─T5Block: 3-6 [2, 100, 512] 2,360,320
│ │ └─T5Block: 3-7 [2, 100, 512] 2,360,320
│ │ └─T5Block: 3-8 [2, 100, 512] 2,360,320
│ └─T5LayerNorm: 2-4 [2, 100, 512] 512
│ └─Dropout: 2-5 [2, 100, 512] --
├─T5Stack: 1-4 [2, 6, 100, 64] 16,449,536
│ └─Embedding: 2-6 [2, 100, 512] (recursive)
│ └─Dropout: 2-7 [2, 100, 512] --
│ └─ModuleList: 2-8 -- --
│ │ └─T5Block: 3-9 [2, 100, 512] 3,147,456
│ │ └─T5Block: 3-10 [2, 100, 512] 3,147,264
│ │ └─T5Block: 3-11 [2, 100, 512] 3,147,264
│ │ └─T5Block: 3-12 [2, 100, 512] 3,147,264
│ │ └─T5Block: 3-13 [2, 100, 512] 3,147,264
│ │ └─T5Block: 3-14 [2, 100, 512] 3,147,264
│ │ └─T5Block: 3-15 [2, 100, 512] 3,147,264
│ │ └─T5Block: 3-16 [2, 100, 512] 3,147,264
│ └─T5LayerNorm: 2-9 [2, 100, 512] 512
│ └─Dropout: 2-10 [2, 100, 512] --
├─Linear: 1-5 [2, 100, 32128] 16,449,536
==============================================================================================================
Total params: 76,961,152
Trainable params: 76,961,152
Non-trainable params: 0
Total mult-adds (M): 186.86
==============================================================================================================
Input size (MB): 0.00
Forward/backward pass size (MB): 217.84
Params size (MB): 307.84
Estimated Total Size (MB): 525.69
==============================================================================================================
13 changes: 13 additions & 0 deletions tests/torchinfo_xl_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -132,3 +132,16 @@ def test_google() -> None:
# Check googlenet in training mode since InceptionAux layers are used in
# forward-prop in train mode but not in eval mode.
summary(google_net, (1, 3, 112, 112), depth=7, mode="train")


def test_huggingface():
from transformers import AutoModelForSeq2SeqLM

model = AutoModelForSeq2SeqLM.from_pretrained("google/flan-t5-small")
inputs = {
"input_ids": torch.zeros(2, 100).long(),
"attention_mask": torch.zeros(2, 100).long(),
"labels": torch.zeros(2, 100).long(),
}
with torch.no_grad():
summary(model, input_data=inputs)
6 changes: 6 additions & 0 deletions torchinfo/layer_info.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -129,6 +129,12 @@ def nested_list_size(inputs: Sequence[Any]) -> tuple[list[int], int]:
size = []
elem_bytes = list(inputs.values())[0].element_size()
for _, output in inputs.items():
# fix for huggingface modules that return tuple
if not isinstance(output, torch.Tensor):
temp = output
while isinstance(temp, tuple):
temp = temp[0]
output = temp
size = list(output.size())
if batch_dim is not None:
size = [size[:batch_dim] + [1] + size[batch_dim + 1 :]]
Expand Down