akashc1

Importing `torch` after importing `decord` causes hanging behavior

6

I ran into a strange issue where the import order specifically between `torch` and `decord` can causing hanging behavior. Concretely, I found that having your imports like this (which could...

llama 3.1 has correct `max_seq_len` for all versions

7

#### Context What is the purpose of this PR? Is it to - [ ] add a new feature - [x] fix a bug - [ ] update tests and/or...

Llama3.1 models do not allow configuring `max_seq_len`

6

Llama 3.1 model builders hardcode the max context length, even though the component builders allow specifying it: - [Llama 3.1 8B](https://github.com/pytorch/torchtune/blob/v0.5.0/torchtune/models/llama3_1/_model_builders.py#L34) - [Llama 3.1 70B](https://github.com/pytorch/torchtune/blob/v0.5.0/torchtune/models/llama3_1/_model_builders.py#L55) - [Llama 3.1 405B](https://github.com/pytorch/torchtune/blob/v0.5.0/torchtune/models/llama3_1/_model_builders.py#L76) -...