App working

Dockerfile

@@ -1,16 +1,13 @@
-# Read the doc: https://huggingface.co/docs/hub/spaces-sdks-docker
-# you will also find guides on how best to write your Dockerfile
+FROM python:3.10
 
-FROM python:3.9
+WORKDIR /app
 
-RUN useradd -m -u 1000 user
-USER user
-ENV PATH="/home/user/.local/bin:$PATH"
+COPY requirements.txt .
 
-WORKDIR /app
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY . .
 
-COPY --chown=user ./requirements.txt requirements.txt
-RUN pip install --no-cache-dir --upgrade -r requirements.txt
+EXPOSE 7860
 
-COPY --chown=user . /app
 CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -1,7 +1,33 @@
 from fastapi import FastAPI
+from pydantic import BaseModel
+import torch
+from transformers import AutoTokenizer, T5ForConditionalGeneration
+
+MODEL_NAME = "google/byt5-small"
 
 app = FastAPI()
 
-@app.get("/")
-def greet_json():
-    return {"Hello": "World!"}
+print("Loading model...")
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = T5ForConditionalGeneration.from_pretrained(MODEL_NAME)
+model.eval()
+
+print("Model loaded.")
+
+class TextRequest(BaseModel):
+    text: str
+
+
+@app.post("/predict")
+def predict(request: TextRequest):
+    prompt = f"Text: {request.text}\nIPA:"
+
+    inputs = tokenizer(prompt, return_tensors="pt", truncation=True, max_length=512)
+
+    with torch.no_grad():
+        outputs = model.generate(**inputs, max_new_tokens=64)
+
+    result = tokenizer.decode(outputs[0], skip_special_tokens=True)
+
+    return {"ipa": result}

requirements.txt

@@ -1,2 +1,7 @@
 fastapi
-uvicorn[standard]
+uvicorn
+torch
+transformers
+sentencepiece
+safetensors
+numpy<2