AliengirlLiv

Currently, the embedding for the encoder is inside the encoder while the embedding for the decoder is inside the sequence to sequence model. The embedding being present in the decoder...

It'd be really helpful to have some way to visualize the a model's execution for both its forward and backward pass. Specifically, an animation where we see the progress of...

The tree to sequence anc model currently decreases its loss by a slight amount from around 8ish to 4/5ish. One potential thing to look at is the values/gradients of operations...

Multi gpu

2

This runs on multiple GPUS. That said, there are some sketchy things: - I just chose num_cpus equal to the number of cpus on my desktop, but IDK what the...

Multi-GPU training errors with peft

1

### 🐛 Describe the bug When I try to use multi-gpu training with accelerate I get an error. Code: ``` import trlx from peft import LoraConfig, TaskType from trlx.data.configs import...

This is a meta-issue for implementing the tutorial [described in the release plan doc](https://docs.google.com/document/d/1WAvjPATtaIQFn1MwzxFf2qA41TkKLacdOWM1JhzpBSk/edit#bookmark=id.hr927kfmt62i). Because this is a pretty chunky task, I'm going to break it up into several smaller...

User feedback that as a new user it's more fun to repeatedly attack accounts that don't just say "access denied" all the time.

So we can recover accounts for users.

- [ ] Gray out the textbox - [ ] Gray out the submit button