{"id":20613009,"url":"https://github.com/experiencor/deep-viz-keras","last_synced_at":"2025-04-15T07:09:47.425Z","repository":{"id":49766887,"uuid":"95198890","full_name":"experiencor/deep-viz-keras","owner":"experiencor","description":"Implementations of some popular Saliency Maps in Keras","archived":false,"fork":false,"pushed_at":"2019-05-11T13:33:01.000Z","size":2482,"stargazers_count":166,"open_issues_count":9,"forks_count":31,"subscribers_count":7,"default_branch":"master","last_synced_at":"2025-04-15T07:09:40.789Z","etag":null,"topics":["convolutional-neural-networks","deep-learning","gradient","keras","machine-learning","saliency-map","visualization"],"latest_commit_sha":null,"homepage":"https://experiencor.github.io/cnn_visual.html","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/experiencor.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2017-06-23T08:16:07.000Z","updated_at":"2025-03-24T00:31:02.000Z","dependencies_parsed_at":"2022-09-22T11:52:55.351Z","dependency_job_id":null,"html_url":"https://github.com/experiencor/deep-viz-keras","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/experiencor%2Fdeep-viz-keras","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/experiencor%2Fdeep-viz-keras/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/experiencor%2Fdeep-viz-keras/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/experiencor%2Fdeep-viz-keras/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/experiencor","download_url":"https://codeload.github.com/experiencor/deep-viz-keras/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":249023723,"owners_count":21199960,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["convolutional-neural-networks","deep-learning","gradient","keras","machine-learning","saliency-map","visualization"],"created_at":"2024-11-16T11:08:39.447Z","updated_at":"2025-04-15T07:09:47.399Z","avatar_url":"https://github.com/experiencor.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"\nThis repository contains the implementations in Keras of various methods to understand the prediction by a Convolutional Neural Networks. Implemented methods are:\n\n* Vanila gradient [https://arxiv.org/abs/1312.6034]\n* Guided backprop [https://arxiv.org/abs/1412.6806]\n* Integrated gradient [https://arxiv.org/abs/1703.01365]\n* Visual backprop [https://arxiv.org/abs/1611.05418]\n\nEach of them is accompanied with the corresponding smoothgrad version [https://arxiv.org/abs/1706.03825], which improves on any baseline method by adding random noise.\n\nCourtesy of https://github.com/tensorflow/saliency and https://github.com/mbojarski/VisualBackProp.\n\n# Examples\n\n* Dog\n\n\u003cimg width=\"700\" src=\"images/doberman_viz.png\"\u003e\n\n* Dog and Cat\n\n\u003cimg width=\"700\" src=\"images/cat_dog_viz.png\"\u003e\n\n\n# Usage\n\ncd deep-viz-keras\n\n```python\nfrom guided_backprop import GuidedBackprop\nfrom utils import *\nfrom keras.applications.vgg16 import VGG16\n\n# Load the pretrained VGG16 model and make the guided backprop operator\nvgg16_model = VGG16(weights='imagenet')\nvgg16_model.compile(loss='categorical_crossentropy', optimizer='adam')\nguided_bprop = GuidedBackprop(vgg16_model)\n\n# Load the image and compute the guided gradient\nimage = load_image('/path/to/image')\nmask = guided_bprop.get_mask(image)               # compute the gradients\nshow_image(mask)                                  # display the grayscaled mask\n```\n\nThe examples.ipynb contains the demos of all implemented methods using the built-in VGG16 model of Keras.\n\n# Notes\n\n+ To compute gradient of any output w.r.t. any input https://github.com/experiencor/deep-viz-keras/issues/5#issuecomment-376452683.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fexperiencor%2Fdeep-viz-keras","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fexperiencor%2Fdeep-viz-keras","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fexperiencor%2Fdeep-viz-keras/lists"}