torchinfo
torchinfo copied to clipboard
View model summaries in PyTorch!
**Describe the bug** `summary` wrongly reports modules as recursive and several entries are duplicated and not in any logical order. **To Reproduce** Steps to reproduce the behavior: 1. `conda create...
**Describe the bug** When the summary on the Graph Convolutional Networks model is called and the model input is provided, the output summary doesn't contain any numbers in the column...
**Describe the bug** `torchinfo.summary()` fails on a model that returns numpy ndarray instead of tensors. **To Reproduce** first install https://mmpose.readthedocs.io ``` from pathlib import Path import mim import torchinfo import...
**Describe the bug** Memory estimation inconsistent with actual GPU memory utilization **To Reproduce** - I am using a simple UNet with 2 layers (same as [here](https://github.com/niuchuangnn/noise2sim/blob/master/noise2sim/modeling/networks/unet.py#L89)). - The input size...
Hi, I find torchinfo seems omit the MAccs of some operations like `torch.mul` and `torch.matmul`. Is it some feature-to-be-added, or we do not need to count these operations when calculating...
Trying to summarize a [CLIP](https://github.com/openai/CLIP) model results in an output where there are no values listed in the `Output Shape` column: ``` ========================================================================================== Layer (type:depth-idx) Output Shape Param # ==========================================================================================...
**Describe the bug** Memory estimates are inconsistent with actual GPU usage for recursive models **To Reproduce** Example code: ```python import torch from torch.nn import Sequential, Linear class MyModel(torch.nn.Module): def __init__(self,...
Related to issue #65
**Describe the bug** A clear and concise description of what the bug is. **To Reproduce** Steps to reproduce the behavior: ``` from transformers import BertModel from torchinfo import summary bert_base_path...
**Describe the bug** When the linear layer has a multidimensional input and output (shape with 3 dimensions or more) the computed multiple adds will be incorrect. **To Reproduce** Add line...