Upload folder using huggingface_hub

README.md

@@ -94,6 +94,9 @@ Three 30.1M parameter models are available in the [checkpoints directory](https:
 - **PER:** 0.49
 - **GER:** 0.39
 
+**Note:** CUPE models are designed for contextless phoneme prediction and are not optimal for phoneme classification tasks that require contextual information. For lower Phoneme Error Rate (PER) metrics, consider using large-scale pretrained models that process entire utterances with contextual awareness. CUPE excels at extracting pure, frame-level embeddings that represent the acoustic properties of each phoneme (e.g., /a/) independently of surrounding context.
+
+
 ---