Adding unit tests for each tuning type #79
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Merge first init
Signed-off-by: Sukriti-Sharma4 <[email protected]>
Signed-off-by: Sukriti-Sharma4 <[email protected]>
…ck/minor_fixes allows disable flash attn and torch dtype param
* first refactor on train Signed-off-by: Sukriti-Sharma4 <[email protected]> * fix:None peft type for fine tuning Signed-off-by: Sukriti-Sharma4 <[email protected]> * accept None as tuningconfig Signed-off-by: Sukriti-Sharma4 <[email protected]> * correct docstrings Signed-off-by: Sukriti-Sharma4 <[email protected]> * refactor of get_hf_peft_config Signed-off-by: Sukriti-Sharma4 <[email protected]> * semantics for python 3.9 Signed-off-by: Sukriti-Sharma4 <[email protected]> * support python 3.9 Signed-off-by: Sukriti-Sharma4 <[email protected]> * check accumulate steps>0 Signed-off-by: Sukriti-Sharma4 <[email protected]> --------- Signed-off-by: Sukriti-Sharma4 <[email protected]>
Signed-off-by: Sukriti-Sharma4 <[email protected]>
…ack/fix_parse_args fix : the way args are passed
…ack/lchu-ibm-patch-1 fix full param tuning
* Add PR template Signed-off-by: ted chang <[email protected]> * Add Makefile Signed-off-by: ted chang <[email protected]> * Update .github/ISSUE_TEMPLATE/user_story.md Co-authored-by: Sukriti Sharma <[email protected]> Signed-off-by: ted chang <[email protected]> * Update .github/pull_request_template.md Co-authored-by: Sukriti Sharma <[email protected]> Signed-off-by: ted chang <[email protected]> --------- Signed-off-by: ted chang <[email protected]> Co-authored-by: Sukriti Sharma <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]> Co-authored-by: Kelly A <[email protected]> Co-authored-by: Anh-Uong <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
* add sample unit test Signed-off-by: ted chang <[email protected]> * Update tests/utils/test_data_type_utils.py Co-authored-by: Sukriti Sharma <[email protected]> Signed-off-by: ted chang <[email protected]> * Update tests/utils/test_data_type_utils.py Co-authored-by: Sukriti Sharma <[email protected]> Signed-off-by: ted chang <[email protected]> * re-raise valueError instead of exit Signed-off-by: ted chang <[email protected]> --------- Signed-off-by: ted chang <[email protected]> Co-authored-by: Sukriti Sharma <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
…ner_image Initial commit for trainer image
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
tharapalanivel
commented
Mar 5, 2024
requirements.txt
Outdated
| @@ -1,7 +1,7 @@ | |||
| numpy | |||
| accelerate>=0.20.3 | |||
| packaging | |||
| transformers>=4.34.1 | |||
| transformers>=4.34.1,<4.38.0 | |||
There was a problem hiding this comment.
Specifying upper bound due to error below based on a recent change in transformers.trainer.py library
self.is_fsdp_xla_v2_enabled = args.fsdp_config["xla_fsdp_v2"]
KeyError: 'xla_fsdp_v2'
Is xla_fsdp_v2 a required field in the config? If so, do we update ours?
There was a problem hiding this comment.
the reason you are seeing this error is because TrainingArguments is not initialized properly. In the code, we inherit from the dataclass, but we did not call .__post_init__. This PR #53 fixes this issue.
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
|
Not sure if its a good idea to commit binary files that are 17MB large into a git repo. |
anhuong
reviewed
Mar 6, 2024
anhuong
reviewed
Mar 6, 2024
tests/helpers.py
Outdated
| from tuning.config import configs, peft_config | ||
|
|
||
|
|
||
| def causal_lm_train_kwargs(train_kwargs): |
There was a problem hiding this comment.
This has come up a few times now, that this code is duplicated in launch_training script and now is duplicated in tests. Feels like it may be worthwhile to refactor the argument parsing into its own method that can be tested separated in unit tests and called in both places.
There was a problem hiding this comment.
Good point @anhuong, I'll move that fn outside of tests/ for now but leave the launch_training.py refactoring for a separate PR
There was a problem hiding this comment.
@anhuong tuning/utils/config_utils.py a good place for this function?
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
Signed-off-by: Thara Palanivel <[email protected]>
tharapalanivel
force-pushed
the
unit_tests
branch
from
f9a703f to
6d110f5
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description of the change
Adding unit tests for
ptandloratuning method using dummy modelRelated issue number
Closes #74
How to verify the PR
Was the PR tested