Update README.md

README.md

@@ -131,7 +131,7 @@ with torch.no_grad():
 ## Training Details
 
 - **Fine-tuning dataset:** 400,000 pairs of anime images and synthetic English captions
-- **Caption generation:** Synthetic captions were produced using Gemma 3, Gemini 2.0 Flash, Llama 4 Maverick, and GPT-4.1.
+- **Caption generation:** Synthetic captions were produced using Llama 4 Maverick.
 - **Task:** Image-to-text, focused on high-quality anime-style descriptions.
 - **Base model:** [HuggingFaceTB/SmolVLM-500M-Base](https://huggingface.co/HuggingFaceTB/SmolVLM-500M-Base)