Update README.md

README.md

@@ -113,7 +113,7 @@ logger = logging.getLogger(__name__)
 
 MODEL_NAME = "nisten/Biggie-SmoLlm-0.15B-Base"
 MAX_LENGTH = 2048
-BATCH_SIZE = 4
+BATCH_SIZE = 9
 LEARNING_RATE = 2e-4
 MAX_STEPS = 3000
 GRADIENT_ACCUMULATION_STEPS = 2
@@ -351,7 +351,7 @@ if __name__ == "__main__":
 ```
 🚀 Now go forth and train, accelerate that code! 
 
-> **Note:** You'll need about 22GB of VRAM. If you have 8GB, change to batch size 1.
+> **Note:** You'll need about 14GB of VRAM. If you have 8GB, change to batch size 4.
 
 Results will appear in `./capybara_finetuned_results`