Botao Chen 52a21ce78f Free up memory after post training finishes (#770) ... ## context Currently, the GPU memory will be continuously occupied after the training finishes. In this PR, we explicitly delete the reference and clean up the memory after training finishes. ## test Before the change, after training a llama 3.2 3B model, >6GB GPU memory is still occupied After the change, after training a llama 3.2 3B model, the GPU memory drops to ~1GB <img width="156" alt="Screenshot 2025-01-14 at 6 05 17 PM" src="https://github.com/user-attachments/assets/45d212b1-a651-49f3-aad9-1c0a27fcebcf" />		2025-01-14 19:19:38 -08:00
..
apis	[post training] define llama stack post training dataset format (#717)	2025-01-14 12:48:49 -08:00
cli	Fix telemetry to work on reinstantiating new lib cli (#761)	2025-01-14 11:31:50 -08:00
distribution	removing unused script file	2025-01-14 17:54:22 -08:00
providers	Free up memory after post training finishes (#770)	2025-01-14 19:19:38 -08:00
scripts	Fix to conda env build script	2024-12-17 12:19:34 -08:00
templates	Fix fireworks run-with-safety template (#766)	2025-01-14 15:28:55 -08:00
__init__.py	export LibraryClient	2024-12-13 12:08:00 -08:00