updates

2026-10-09 00:09:45 +00:00 · 2024-11-29 10:52:12 +05:30
parent ced8558eb0
commit 4e3bb7ab5f
1 changed files with 1 additions and 0 deletions
@@ -89,6 +89,7 @@ Our script currently doesn't leverage `accelerate` and some of its consequences
 * No support for distributed training. 
 * No intermediate checkpoint saving and loading support.
 * `train_batch_size > 1` are supported but can potentially lead to OOMs because we currently don't have gradient accumulation support.
+* No support for 8bit optimizers (but should be relatively easy to add).

 **Misc**: