Update finetune_llama3.py

marcopoli · May 11, 2024 · eaf49a6 · eaf49a6
1 parent 1704d0a
commit eaf49a6
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/model_adaptation/finetune_llama3.py b/model_adaptation/finetune_llama3.py
@@ -12,7 +12,7 @@
 
 max_seq_length = 8192 # Choose any! We auto support RoPE Scaling internally!
 dtype = None # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+
-load_in_4bit = False # Use 4bit quantization to reduce memory usage. Can be False.
+load_in_4bit = True # Use 4bit quantization to reduce memory usage. Can be False.
 model_name = "swap-uniba/LLaMAntino-3-ANITA-8B-Inst-DPO-ITA"
 
 model, tokenizer = FastLanguageModel.from_pretrained(