-
Notifications
You must be signed in to change notification settings - Fork 15
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
hydra_zen.hydrated_dataclass
unconditionally sets the docstring on the wrapped class using the target's __doc__
#750
Comments
lebrice
added a commit
to mila-iqia/ResearchTemplate
that referenced
this issue
Nov 12, 2024
- mit-ll-responsible-ai/hydra-zen#750 Signed-off-by: Fabrice Normandin <[email protected]>
lebrice
added a commit
to mila-iqia/ResearchTemplate
that referenced
this issue
Nov 15, 2024
* WIP: Add an LLM finetuning example Signed-off-by: Fabrice Normandin <[email protected]> * WIP: add / rename more configs Signed-off-by: Fabrice Normandin <[email protected]> * Finetuning example seems to be working Signed-off-by: Fabrice Normandin <[email protected]> * Making progress, more self-contained example Signed-off-by: Fabrice Normandin <[email protected]> * Works! (need to fix the hash used for path though) Signed-off-by: Fabrice Normandin <[email protected]> * Improve hashing, reduce default block size Signed-off-by: Fabrice Normandin <[email protected]> * Fix val_loss logging and add docstring Signed-off-by: Fabrice Normandin <[email protected]> * Increase the number of dataloader workers Signed-off-by: Fabrice Normandin <[email protected]> * Use smaller model for now Signed-off-by: Fabrice Normandin <[email protected]> * Use FSDP in the example Signed-off-by: Fabrice Normandin <[email protected]> * Fix bug in id generation from config classes Signed-off-by: Fabrice Normandin <[email protected]> * Tweak config, try to setup mid-epoch checkpointing Signed-off-by: Fabrice Normandin <[email protected]> * Rename `HFExample` -> `TextClassificationExample` Signed-off-by: Fabrice Normandin <[email protected]> * Fix broken links in nav Signed-off-by: Fabrice Normandin <[email protected]> * Remove "huggingface" datamodule config Signed-off-by: Fabrice Normandin <[email protected]> * Fix issues in config/tests for text_classification Signed-off-by: Fabrice Normandin <[email protected]> * Add an entry to test the llm_finetuning_example Signed-off-by: Fabrice Normandin <[email protected]> * Fix issues in the text classification example Signed-off-by: Fabrice Normandin <[email protected]> * Fix weird docstring issues with hydra-zen - mit-ll-responsible-ai/hydra-zen#750 Signed-off-by: Fabrice Normandin <[email protected]> * Fix test and config of text_classification_example Signed-off-by: Fabrice Normandin <[email protected]> * Move test from main_test.py to example_test.py Signed-off-by: Fabrice Normandin <[email protected]> * forward_pass is a method of LearningAlgorithmTests Signed-off-by: Fabrice Normandin <[email protected]> * Various type hint fixes and tweaks Signed-off-by: Fabrice Normandin <[email protected]> * WIP: Adding some tests for LLM finetuning example Signed-off-by: Fabrice Normandin <[email protected]> * Fix issue in `jax.md` Signed-off-by: Fabrice Normandin <[email protected]> * Add link to the example page in index.md Signed-off-by: Fabrice Normandin <[email protected]> * Fix tests for the llm finetuning example Signed-off-by: Fabrice Normandin <[email protected]> * Fix issue with tuples in regression files Signed-off-by: Fabrice Normandin <[email protected]> * Fix test for `get_hash_of` Signed-off-by: Fabrice Normandin <[email protected]> * Remove unused _field function Signed-off-by: Fabrice Normandin <[email protected]> * Fix issue with built-in modules in autoref plugin Signed-off-by: Fabrice Normandin <[email protected]> * Add a bit of info in the example doc Signed-off-by: Fabrice Normandin <[email protected]> * Add more links in the doc of the module Signed-off-by: Fabrice Normandin <[email protected]> * Fix issue with the text classification example Signed-off-by: Fabrice Normandin <[email protected]> * Add skipif mark for LLM finetuning test Signed-off-by: Fabrice Normandin <[email protected]> * Fix data_dir of text_classification_example Signed-off-by: Fabrice Normandin <[email protected]> * Use the "auto" strategy for LLM Finetuning tests Signed-off-by: Fabrice Normandin <[email protected]> * Fix error in fork_rng of LLM finetuning example Signed-off-by: Fabrice Normandin <[email protected]> * Try a hacky fix for failing test Signed-off-by: Fabrice Normandin <[email protected]> * Don't run llm finetuning tests on github Cloud CI Signed-off-by: Fabrice Normandin <[email protected]> * Add missing regression files Signed-off-by: Fabrice Normandin <[email protected]> * Rename llm_finetuning_example -> llm_finetuning Signed-off-by: Fabrice Normandin <[email protected]> * Fix import error Signed-off-by: Fabrice Normandin <[email protected]> --------- Signed-off-by: Fabrice Normandin <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi there!
When you use
hydrated_dataclass
, the__doc__
on the wrapped dataclass is unconditionally set to "A structured config designed to ..." followed by the docstring from the_target_
.This causes issues with doctest, since it picks up the doctests from the docstring of the target callable!
An example:
The text was updated successfully, but these errors were encountered: