Deep Learning in Medical Imaging: Multi-label Classification with PyTorch | Hands-on Demo

[@vlad910]

Thank you, sir. I wonder how you handled the data imbalance that you mentioned on your blog post.

[@nitikasaun8689]

Sir I am working on waste localization and seggregation using image dataset. From the input image multiple labels to give the output of its presence, which model is best suited for this also TO improve accuracy what can be done?

[@iyshwaryakannan6677]

Hi sir, thank you for your insights. If we want to train and detect various attributes of a person, like his gender, his attire colour, and his attire type. what kind of training can be used sir

[@LearnOpenCV]

It depends on the problem statement and datasets available. If you have a dataset that differentiates between genders and multiple attires, you can train a classification model of your choice.

[@J4BEZ]

I am very appreciate for your hard effort to teach us what is multi-label classification and how can we train the pytorch model with custom dataset🙇‍♂🙇‍♂ But now, there seems to be a liitle issue with package installation dependency how can I solve the problem and have a experience with this great course😥

[@LearnOpenCV]

Hello, we ran the notebooks and encountered this error: 'ImportError: cannot import name 'ModelMetaclass' from 'pydantic.main' (/usr/local/lib/python3.10/dist-packages/pydantic/main.py)' If just in case you are also facing the same issue on google Colab, we simply need to install the latest packages for pytorch-lightning. The lightning team had fixed the issue in their later releases. Please the make the following changes: Update the first "pip install" packages line to: "!pip install -qqqU kaggle torchinfo lightning torchmetrics. P.N: If this isn't the error you are facing, can you please paste the error log message here?

[@J4BEZ]

@@LearnOpenCV Thank you for your consideration🙇‍♂🙇‍♂ I 'll try it as soon as possible

[@J4BEZ]

​@@LearnOpenCV when I changed first line, now I can find the package dependency problem of lightning was solved! but there is a little issue beside :: ``` ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. lida 0.0.10 requires kaleido, which is not installed. llmx 0.0.15a0 requires cohere, which is not installed. llmx 0.0.15a0 requires openai, which is not installed. llmx 0.0.15a0 requires tiktoken, which is not installed. tensorflow-probability 0.22.0 requires typing-extensions

[@J4BEZ]

​@@LearnOpenCV Oh..! I'm very sorry for late. I think my comment was blocked that considered as spam😥😥 When I change the first line "!pip install -qqqU kaggle.. " to "!pip install -qqqU kaggle torchinfo lightning torchmetrics", the dependecy problem of torchmetrics has disappeared!! then I got an issue that said " ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. lida 0.0.10 requires kaleido, which is not installed." so I added a new code block with "!pip install kaleido" below the "1 Install & Import Necessary Libraries", the package installation problem has cleared!! Thank you very much😆😆

[@LearnOpenCV]

Glad that it worked out for you!@@J4BEZ