Captum is a PyTorch library for model interpretability. It provides algorithms and tools to understand and attribute the predictions of PyTorch models. It supports various modalities including vision and text, and is built to be extensible for interpretability research. Key features include support for multi-modal models, PyTorch integration, and an open-source, generic library design.
• built on pytorch
• multi-modal support
• extensible
The purpose of target is to select a single (scalar) value for each example in the output of your model to compute attributions based on the given target parameter.
To address this issue, you can either reduce n_steps, which may lead to lower-quality approximations, or use the internal_batch_size argument, which allows dividing the expanded input into batches.
To enable this, simply set the perturbations_per_eval argument to the desired value. If you have multiple GPUs machines available, you can also wrap your model with DataParallel.
Yes! SmoothGrad and VarGrad are available through NoiseTunnel in Captum, which can be used with any attribution algorithm in Captum. More details on Noise Tunnel can be found in the documentation.
We have a tutorial demonstrating usage of Integrated Gradients on BERT. For NLP models that take token indices as inputs, we cannot take gradients with respect to indices.
To apply gradient-based attribution methods, it is necessary to replace the embedding layer with an InterpretableEmbedding layer or use LayerIntegratedGradients to compute attribution with respect to the embedding output.
Most methods will work fine with functional non-linearities and arbitrary operations. Some methods, which require placing hooks during back-propagation, will not work appropriately.
Yes, we have support for all these model types. Note that JIT models do not yet support hooks, so any methods using hooks including layer and neuron attribution methods are not supported.
New attribution algorithms that fit the structure of Captum can be considered for contribution to the contrib package of algorithms in Captum.
To resolve the issue you can set`torch.backends.cudnn.enabled` flag to False.
There are currently no job postings for this AI tool.
Average Rating: 0.0
5 Stars:
0 Ratings
4 Stars:
0 Ratings
3 Stars:
0 Ratings
2 Stars:
0 Ratings
1 Star:
0 Ratings
No ratings available.
Xjoy.ai provides AI tools for photo editing, face swapping, pose generation, short video creation, and dance animation.
View DetailsAngel.ai powers immersive experiences with AI Angels. Chat with AI girlfriends and boyfriends, generate images, and create personalized AI companions.
View DetailsConnect your Github repos to ChatGPT & Claude for code assistance, bug finding, and documentation. Free trial available.
View DetailsIncite AI is an AI-powered platform providing real-time intelligence and analysis for financial markets, including stocks, crypto, and ETFs. It offers insights, predictions, and tools for informed investment decisions.
View DetailsGatsbi AI is an AI co-scientist that crafts tailored solutions for research challenges and generates publication-ready papers and patent documents effortlessly, supporting ideation, scholarly writing, and patent writing.
View DetailsSprunky is an interactive music game where players create tunes by mixing beats, effects, and vocals with unique characters. A fan-made modification of Incredibox for creative music composition.
View DetailsA trivia website with questions in multiple categories. Play now and expand your knowledge!
View DetailsThryveChat is a free 24/7 AI-powered companion for emotional, wellbeing & fitness support. Chat freely, get support, daily reminders & inspiration to start your journey to a healthier, happier you today. Thryve speaks 50+ languages!
View Details