salman

Write docpages for #1005 and the various losses supported for our DPO recipe

[RFC][Docs][WIP] Custom Recipe Tutorial

5

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

Toggling KV-caches

7

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

We'll probably also need #1782.

https://github.com/pytorch/torchtune/issues/1795

[RFC] Classification task fine-tuning

9

There has been some community appetite for classification tasks #1249 #1124. Incidentally, due to the use of classification models for RLHF, we already have some of the necessary components to...

Now that we've deprecated these classes we can remove them in an upcoming release.

[RFC] Supporting KV-cache toggling

9

### Problem Currently, when we use `model.setup_caches()`, KV-caches are always updated for every subsequent forward pass on the model. We have valid use cases for using `model.setup_caches()`, but then not...

[WIP] Config Continous Integration (CCI)

6

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

Several of our tests fail due to precision differences between MPS and CUDA/CPU falling outside of numerical tolerances on our tests: ```bash $ pytest tests FAILED tests/torchtune/generation/test_generation.py::TestGenerate::test_stop_tokens_batched[prompt_tokens_batched-generation_model_no_kv_cache] - assert False...