Update README.md

README.md

@@ -103,9 +103,7 @@ This model can be converted to GGML format for use with [whisper.cpp](https://gi
 
 ## Limitations
 
-- Trained primarily on read speech (Common Voice) and formal speech (VoxPopuli) — may underperform on casual/noisy audio (e.g., phone calls, WhatsApp voice messages)
-- Whisper-tiny has limited capacity (39M params); for better accuracy consider whisper-small or whisper-medium
-- The multilingual tokenizer is suboptimal for Italian-only use (apostrophe-heavy grammar pays a token tax)
+- WER is not as good as whisper tiny.en, also italian mixed with english like terms tends to fail at the moment
 
 ## License