-
I have been observing low cosine similarity scores for InternVideo2 video embeddings compared to relevant text caption embeddings. In some cases, the scores are even negative. I am not sure if I am mi…
-
Description: I have tested a set of images using your model weights and performed N
matching based on cosine similarity. I followed the code structure below:
# Sample code for N:N matching using…
-
Thanks for the great work. I notice that in the Attention and FFN, the output matrix (i.e., self.to_out) is normalized differently along the first dimension instead of the last dimension (normalizing …
-
Currently, the filter apodization options is:
```js
const apodization2DOptions = {
sineBell: {
apply: true, //false if the shape should be ignored
shape: {
kind: 'sineBell',
opt…
-
Most of the sine tables do not start at the origin, but rather at -128. This is bad because it can cause a click if the wave is played raw. They should start at 0, then increase from there, then bac…
eclab updated
2 weeks ago
-
Hi,
I get the following error when I was performing the quick run on MPRC:
ModuleNotFoundError: No module named 'trainers.hooks.cosine_lr_scheduler'
I looked manually, and cosine_lr_scheduler d…
-
I tried using dog.pt. For any two dogs, the cosine similarity of their outputs is always greater than 0.99. Do you know what might be the issue?
-
Hi,
I'm trying to separate the two legends to the right such that the top one is closer to the middle and the bottom one is closer to the bottom plot:
> Error in tcrossprod(x = rescale(x), y = rescale(y)) : n…