Create requirements.txt

requirements.txt

@@ -0,0 +1,9 @@
+# requirements.txt ── deps for Hugging Face Space
+streamlit>=1.33         # front-end UI
+transformers>=4.40      # Hugging Face pipelines
+torch>=2.2              # core tensor library (CPU build auto-selected by HF)
+soundfile               # write WAV audio
+numpy                   # array ops for audio concat
+pillow                  # image handling (PIL)
+sentencepiece           # tokenizer back-end for some models
+accelerate              # speeds up HF inference, optional but lightweight