Remove max_length from tokenization (meta-llama#604)

mreso · web-flow · commit 01a20d1e86f8 · 2024-07-23T17:38:00.000+02:00
diff --git a/recipes/quickstart/inference/local_inference/inference.py b/recipes/quickstart/inference/local_inference/inference.py
@@ -89,7 +89,6 @@ def inference(
 
         batch = tokenizer(
             user_prompt,
-            padding="max_length",
             truncation=True,
             max_length=max_padding_length,
             return_tensors="pt",
