-
Add support for identifying a subset of the input tokens that is most influential on the sentiment prediction (a salience map, rationale or explanation).
- [x] Check bertology survey for list of me…
-
Obtain I/O tags from rationales with (globally) optimal relative rationale length and repeat experiment shown in Fig 1 with this annotation instead of SEA.
Caveats:
* If saliency is strongly corre…
-
Hi Shangran Qiu
Thank you for providing code for your Brain 2020 paper, but I have the following questions.
1. Random selection 47*47*47 patches can ensure the distinction of features? If AD and NC …
-
## 🐛 Bug
I was trying to apply Integrated Gradients to the detectron2 implementation of a COCO pre-trained Faster RCNN (specifically COCO-Detection/faster_rcnn_R_50_FPN_3x.yaml). In order to get a …
-
Hey
I am trying to use Bertviz to debug my BART model (specifically your notebook https://github.com/jessevig/bertviz/blob/master/notebooks/head_view_bart.ipynb ). I wonder if you have any advice o…
-
Methods:
* fixed relative length, e.g. 20% of input tokens (but at least 1 token / always rounding up?)
* fixed absolute length rationale, e.g. 5 (shorter if input is shorter than 5 tokens)
* thres…
-
Hi, I implemented the following CNN model in PyTorch and was testing the different model explanation methods in captum. Most of the methods seem to be working fine but when I tried gradcam using the …
-
Hi,
I am currently applying Saliency to interpret the results of my custom Bert model, which outputs a single non-normalised value per example (no softmax or sigmoid applied) for a binary classific…
-
Hello, sir thank you once again
I read the paper with code and I deeply study the model with important changes in your model. I am confused as mentioned AUC is 0.82 but how I can get this I mean in …
-
Can you please provide an end-to-end example of how to run AllenNLP interpret on some custom text input? Thanks!