Spaces:

Mdrnfox
/

peft-bench-eval

Sleeping

Mdrnfox commited on 28 days ago

Commit

09122b5

verified ·

1 Parent(s): b3b0a24

Update run_eval.py

Files changed (1) hide show

run_eval.py CHANGED Viewed

@@ -45,7 +45,7 @@ for cfg in CONFIGS:
     tasks = cfg["tasks"]
     print(f"\nLoading base model: {base_model_id}")
-    tokenizer = AutoTokenizer.from_pretrained(base_model_id, use_fast=True)
     # Try causal first, fallback to encoder
     try:

     tasks = cfg["tasks"]
     print(f"\nLoading base model: {base_model_id}")
+    tokenizer = AutoTokenizer.from_pretrained(base_model_id, use_fast=False)
     # Try causal first, fallback to encoder
     try: