-
paper: https://arxiv.org/pdf/2410.05258
![キャプチャ](https://github.com/user-attachments/assets/c3b759e2-2fb4-465a-bad5-8675593e5518)
Diff attention part:
https://github.com/microsoft/unilm/blob/mast…
-
## Descirbe the bug
Hi, I want to do multihead finetuning on personal pre-trained model(`ptbp_model.model`based on version 0.3.7, main branch), after editing these commands
- `--foundation_model='..…
-
Hi, thank you for the great baseline repo.
I am trying to setup each dataset (with different number of classes) as a task and perform continual learning. However I am a little lost regarding how the…
-
I can't figure out how to get keynav to work in a multihead environment. I use it with XMonad, where I can switch to a head, but the monitor selection isn't clear to me (there are other tools that als…
-
There is a great need in models with multiple heads. For instance, classification and segmentation.
-
Hello,
Would it be possible to cycle only in the tabs of one tabbed container?
Currently, with multiple screens connected, the focus leaves the tabbed container and goes to the next monitor.
…
-
Should be just a matter of modifying `calculate_geometry` in `src/bar_builder.rs` to take multiple displays into account and giving the user control over which monitor to display the bar on.
Displa…
-
Hi, thanks for reproducing Differential Transformer. It seems there are some problems in your reproducing code. You should split q and k in n_head dimension, do re-parameterization for lambda, and add…
-
```
What steps will reproduce the problem?
1. Get an ATI card and configure the fglrx driver
2. Get "big desktop" working with XGL and Beryl
3. Start AWN
What is the expected output? What do you see …
-
I find the matrix operations in [Chpt. 16](https://github.com/rasbt/machine-learning-book/blob/94785477fdabd83473de6300d4fa5e50e89b9684/ch16/ch16-part1-self-attention.ipynb) confusing. For example, in…