diff --git a/unsloth/models/llama.py b/unsloth/models/llama.py
index b5bfa3cbfa..7bee733a15 100644
--- a/unsloth/models/llama.py
+++ b/unsloth/models/llama.py
@@ -1882,8 +1882,8 @@ class FastLlamaModel:
         f"==((====))==  Unsloth - 2x faster free finetuning | Num GPUs used = {len(set(p.device for p in model.parameters()))}\\n"\\
         f"   {chr(92)}{chr(92)}   /|    Num examples = {num_examples:,} | Num Epochs = {num_train_epochs:,} | Total steps = {max_steps:,}\\n"\\
         f"O^O/ {chr(92)}_/ {chr(92)}    Batch size per device = {self._train_batch_size:,} | Gradient accumulation steps = {args.gradient_accumulation_steps}\\n"\\
-        f"{chr(92)}        /    Data Parallel GPUs = {args.world_size} | Total batch size = {total_train_batch_size:,}\\n"\\
-        f' "-____-"     Number of trainable parameters = {get_model_param_count(model, trainable_only=True):,}/{get_model_param_count(model)}'
+        f"{chr(92)}        /    Data Parallel GPUs = {args.world_size} | Total batch size ({self._train_batch_size}*{args.gradient_accumulation_steps}*{args.world_size}) = {total_train_batch_size:,}\\n"\\
+        f' "-____-"     Trainable parameters = {get_model_param_count(model, trainable_only=True):,}/{get_model_param_count(model):,} ({get_model_param_count(model, trainable_only=True)/get_model_param_count(model)*100:.2f})'
         logger.warning(debug_info)
         import subprocess, re, gc
         for _ in range(3):