Add application file

Dockerfile

@@ -0,0 +1,18 @@
+FROM python:3.12-slim
+
+RUN apt-get update && apt-get install -y \
+    libsndfile1-dev \
+    build-essential \
+    libopenblas-dev \
+    && rm -rf /var/lib/apt/lists/*
+    
+WORKDIR /app
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY . .
+
+EXPOSE 8000
+
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "8000", "--log-level", "debug"]

app.py

@@ -0,0 +1,28 @@
+import gradio as gr
+from transformers import pipeline
+
+
+pipe = pipeline(model="pollitoconpapass/whisper-small-finetuned") 
+
+def transcribe(audio):
+    text = pipe(audio)["text"]
+    return text
+
+iface = gr.Interface(
+    fn=transcribe,
+    inputs=gr.Audio(type="filepath"),
+    outputs="text",
+    title="Whisper Small Demo - ZLTech",
+    description='''
+                Realtime demo of speech recognition fine-tuned using Whisper small model. New implementation: Quechua language.
+
+                
+                If you want to use this as an endpoint, go to endpoint.py. 
+
+                Source: https://huggingface.co/pollitoconpapass/whisper-small-finetuned
+
+
+                '''
+)
+
+iface.launch()

docker-compose.yml

@@ -0,0 +1,6 @@
+version: '3.12'
+services:
+  whisper-api-dev:
+    build: .
+    ports:
+      - "8000:8000"

endpoint.py

@@ -0,0 +1,20 @@
+import io
+import uvicorn
+import librosa
+from transformers import pipeline
+from fastapi import FastAPI, File, UploadFile
+
+
+app = FastAPI()
+pipe = pipeline(model="pollitoconpapass/whisper-small-finetuned")
+
+@app.post("/transcribe-whisper")
+async def transcribe(audio: UploadFile = File(...)):
+    contents = await audio.read()
+    buffer = io.BytesIO(contents)
+    with buffer:
+        audio_array, _= librosa.load(buffer, sr=16000)
+
+    text = pipe(audio_array)["text"]
+    return {"text": text}
+

requirements.txt

@@ -0,0 +1,7 @@
+uvicorn==0.25.0
+librosa==0.10.1
+transformers==4.39.0
+fastapi==0.108.0
+soundfile==0.12.1
+torch==2.3.0
+python-multipart