-
## ❓ Questions and Help
Thank your for the great library and your efforts towards explainable AI.
I am writing to find out how the library can be used to better understand a cosine similarity met…
-
## 🐛 Bug
When running the example script `python -m captum.insights.example`, the interpreter raised an exception KeyError: 'WERKZEUG_SERVER_FD'
## To Reproduce
I installed the following de…
-
## 🚀 Feature
Current usage assumes that mdules must output a pytorch tensor and not a tuple: many modules in `transformers` library return multiple outputs, making captum not work with them (e.g. Lla…
-
Hi,
I'm currently working on a sequence tagging problem, and I'm interested in investigating what tokens in the sequence are relevant for each step's predicting.
-
### 🐛 Describe the bug
When attempting to use `CaptumExplainer` to explain a graph-level prediction from a Heterogeneous GNN, I get the following traceback:
```
explanation = explainer(
Fil…
-
## 🐛 Bug
When I run captum insight `python -m captum.insights.example` I get below error
```
KeyError: 'WERKZEUG_SERVER_FD'
```
## To Reproduce
Steps to reproduce the behavior:
Run …
-
### 🐛 Describe the bug
Hi! The new explainer interface fails when we have more than one relation type with the FiLMConv layer:
```
from torch_geometric.explain import Explainer, CaptumExplainer
…
-
## Feature
Support feature importance in tabular data scenarios.
1. Understand which features are beneficial for prediction and help to develop new features
2. Feature selection, removing feature…
-
Hi, I'm trying to use Integrated Gradients on a simple DQN model in my MacBook using the MPS backend.
`model = IntegratedGradients(model)`
`attribution = xai.attribute(torch.tensor(ob, dtype = to…
-
## 🐛 Bug
I'm experiencing RAM leaks when calculating word attributions through https://github.com/cdpierse/transformers-interpret library which delegates the most of the heavy lifting to Captum. So…