foundation-model-stack / fms-hf-tuning

🚀 Collection of tuning recipes with HuggingFace SFTTrainer and PyTorch FSDP.
Apache License 2.0
22 stars 41 forks source link

bug: In the trainer controller, change operations constructor to accept operation name and kwargs #189

Closed seshapad closed 2 months ago

seshapad commented 3 months ago

Describe the bug

The trainer controller has operation classes which are derived from a base operation class. The derived classes are not initialized with name of the operation name and , variable argument kwargs . Furthermore, these are not propagated to the base class constructor.

Platform

Please provide details about the environment you are using, including the following:

Sample Code

NA

Expected behavior

NA

Observed behavior

NA

Additional context

NA