tangledgroup
/

tangled-llama-t-128k-base-v0.1

Text Generation

Inference Endpoints

Model card Files Files and versions Community

mtasic85 commited on 24 days ago

Commit

e03ba1c

•

1 Parent(s): 05227a3

pretrain model

Files changed (1) hide show

scripts/pretrain-model.yaml +2 -2

scripts/pretrain-model.yaml CHANGED Viewed

@@ -114,8 +114,8 @@ eval:
 # Optimizer-related arguments
 optimizer:
   # class_path: torch.optim.AdamW
-  class_path: grokadamw.GrokAdamW
-  # class_path: bitsandbytes.optim.AdamW8bit
   # class_path: bitsandbytes.optim.PagedAdamW8bit
   init_args:

 # Optimizer-related arguments
 optimizer:
   # class_path: torch.optim.AdamW
+  # class_path: grokadamw.GrokAdamW
+  class_path: bitsandbytes.optim.AdamW8bit
   # class_path: bitsandbytes.optim.PagedAdamW8bit
   init_args: