| Name | cka-pytorch JSON |
| Version |
1.1.2
JSON |
| download |
| home_page | None |
| Summary | A PyTorch implementation of Centered Kernel Alignment (CKA) with GPU support. |
| upload_time | 2025-09-17 13:38:21 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.11 |
| license | None |
| keywords |
pytorch
cka
centered kernel alignment
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# Centered Kernel Alignment (CKA) - PyTorch Implementation
A PyTorch implementation of Centered Kernel Alignment (CKA) with GPU support for fast and efficient computation.
> [!WARNING]
> This project is for educational and academic purposes (and for fun 🤷🏻).
## Features
- **GPU Accelerated:** Leverages the power of GPUs for significantly faster CKA calculations compared to NumPy-based implementations.
- **On-the-Fly Calculation:** Computes CKA on-the-fly using mini-batches, avoiding the need to cache large intermediate feature representations.
- **Easy to Use:** Simple and intuitive API for calculating the CKA matrix between two models.
- **Flexible:** Can be used with any PyTorch models and dataloaders.
## Installation
```bash
pip install cka-pytorch
```
## Usage
```python
import torch
from torchvision.models import resnet18
from torch.utils.data import DataLoader
from cka_pytorch.cka import CKACalculator
# 1. Define your models and dataloader
model1 = resnet18(pretrained=True).cuda()
model2 = resnet18(pretrained=True).cuda() # Or a different model
# Create a dummy dataloader for demonstration
dummy_data = torch.randn(100, 3, 224, 224)
dummy_labels = torch.randint(0, 10, (100,))
dummy_dataset = torch.utils.data.TensorDataset(dummy_data, dummy_labels)
dataloader = DataLoader(dummy_dataset, batch_size=32)
# 2. Initialize the CKACalculator
# By default, we will calculate CKA across all layers of the two models
calculator = CKACalculator(
model1=model1,
model2=model2,
model1_name="ResNet18",
model2_name="ResNet18",
batched_feature_size=256,
verbose=True,
)
# 3. Calculate the CKA matrix
cka_matrix = calculator.calculate_cka_matrix(dataloader)
# 4. Plot the CKA Matrix as heatmap
calculator.plot_cka_matrix(title="CKA between ResNet18 and ResNet18")
```
## Contributing
- If you find this repository helpful, please give it a :star:.
- If you encounter any bugs or have suggestions for improvements, feel free to open an issue.
- This implementation has been primarily tested with ResNet architectures.
## Acknowledgement
This project is based on:
- [CKA.pytorch](https://github.com/numpee/CKA.pytorch)
- [centered-kernel-alignment](https://github.com/RistoAle97/centered-kernel-alignment)
Raw data
{
"_id": null,
"home_page": null,
"name": "cka-pytorch",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.11",
"maintainer_email": null,
"keywords": "pytorch, cka, centered kernel alignment",
"author": null,
"author_email": "Dat-Thinh Nguyen <datthinh1801@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/94/fa/da73b1c93025a88c4779756a2967dad4270284fde1c22679e978cb5bf34b/cka_pytorch-1.1.2.tar.gz",
"platform": null,
"description": "# Centered Kernel Alignment (CKA) - PyTorch Implementation\n\nA PyTorch implementation of Centered Kernel Alignment (CKA) with GPU support for fast and efficient computation.\n\n> [!WARNING]\n> This project is for educational and academic purposes (and for fun \ud83e\udd37\ud83c\udffb).\n\n## Features\n\n- **GPU Accelerated:** Leverages the power of GPUs for significantly faster CKA calculations compared to NumPy-based implementations.\n- **On-the-Fly Calculation:** Computes CKA on-the-fly using mini-batches, avoiding the need to cache large intermediate feature representations.\n- **Easy to Use:** Simple and intuitive API for calculating the CKA matrix between two models.\n- **Flexible:** Can be used with any PyTorch models and dataloaders.\n\n## Installation\n```bash\npip install cka-pytorch\n```\n\n## Usage\n\n```python\nimport torch\n\nfrom torchvision.models import resnet18\nfrom torch.utils.data import DataLoader\n\nfrom cka_pytorch.cka import CKACalculator\n\n\n# 1. Define your models and dataloader\nmodel1 = resnet18(pretrained=True).cuda()\nmodel2 = resnet18(pretrained=True).cuda() # Or a different model\n\n# Create a dummy dataloader for demonstration\ndummy_data = torch.randn(100, 3, 224, 224)\ndummy_labels = torch.randint(0, 10, (100,))\ndummy_dataset = torch.utils.data.TensorDataset(dummy_data, dummy_labels)\ndataloader = DataLoader(dummy_dataset, batch_size=32)\n\n# 2. Initialize the CKACalculator\n# By default, we will calculate CKA across all layers of the two models\ncalculator = CKACalculator(\n model1=model1,\n model2=model2,\n model1_name=\"ResNet18\",\n model2_name=\"ResNet18\",\n batched_feature_size=256,\n verbose=True,\n)\n\n# 3. Calculate the CKA matrix\ncka_matrix = calculator.calculate_cka_matrix(dataloader)\n\n# 4. Plot the CKA Matrix as heatmap\ncalculator.plot_cka_matrix(title=\"CKA between ResNet18 and ResNet18\")\n```\n\n## Contributing\n\n- If you find this repository helpful, please give it a :star:.\n- If you encounter any bugs or have suggestions for improvements, feel free to open an issue.\n- This implementation has been primarily tested with ResNet architectures.\n\n## Acknowledgement\nThis project is based on:\n- [CKA.pytorch](https://github.com/numpee/CKA.pytorch)\n- [centered-kernel-alignment](https://github.com/RistoAle97/centered-kernel-alignment)\n",
"bugtrack_url": null,
"license": null,
"summary": "A PyTorch implementation of Centered Kernel Alignment (CKA) with GPU support.",
"version": "1.1.2",
"project_urls": {
"Homepage": "https://github.com/datthinh1801/CKA.pytorch",
"Issues": "https://github.com/datthinh1801/CKA.pytorch/issues",
"Repository": "https://github.com/datthinh1801/CKA.pytorch.git"
},
"split_keywords": [
"pytorch",
" cka",
" centered kernel alignment"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "c50fcabf2db9a66e66ffffe050004bd49c450d3fb7f2e82462a5481a211e5f4b",
"md5": "51bf039e2cc05ef8e5d48ab948dc6ae8",
"sha256": "d2ad1fb911ae293bf9279e153417e9c7b9846e09376d38b7139a58566b284b19"
},
"downloads": -1,
"filename": "cka_pytorch-1.1.2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "51bf039e2cc05ef8e5d48ab948dc6ae8",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.11",
"size": 18206,
"upload_time": "2025-09-17T13:38:19",
"upload_time_iso_8601": "2025-09-17T13:38:19.752372Z",
"url": "https://files.pythonhosted.org/packages/c5/0f/cabf2db9a66e66ffffe050004bd49c450d3fb7f2e82462a5481a211e5f4b/cka_pytorch-1.1.2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "94fada73b1c93025a88c4779756a2967dad4270284fde1c22679e978cb5bf34b",
"md5": "329f6726861210ff7073498a6399771b",
"sha256": "87e267b6d0fb417db3fb6261f0cfafe3cb7c75b370fecac84154efd4931e9af6"
},
"downloads": -1,
"filename": "cka_pytorch-1.1.2.tar.gz",
"has_sig": false,
"md5_digest": "329f6726861210ff7073498a6399771b",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.11",
"size": 16873,
"upload_time": "2025-09-17T13:38:21",
"upload_time_iso_8601": "2025-09-17T13:38:21.024242Z",
"url": "https://files.pythonhosted.org/packages/94/fa/da73b1c93025a88c4779756a2967dad4270284fde1c22679e978cb5bf34b/cka_pytorch-1.1.2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-09-17 13:38:21",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "datthinh1801",
"github_project": "CKA.pytorch",
"github_not_found": true,
"lcname": "cka-pytorch"
}
