update for hf

.dockerignore

@@ -0,0 +1,11 @@
+__pycache__/
+train/
+*.csv
+*.bin
+*.pt
+logs/
+*.ipynb
+*.md
+*.safatensors
+app/finetuned_model/
+finetuned_model/checkpoint-*

Dockerfile

@@ -0,0 +1,11 @@
+FROM python:3.10-slim
+
+WORKDIR /code
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY . /code
+
+EXPOSE 7860
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860"]

app/filter_review.py

@@ -0,0 +1,22 @@
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+import torch
+import torch.nn.functional as F
+
+MODEL_FINETUNED = "yondikavl/artour-spam-filter"
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+model = AutoModelForSequenceClassification.from_pretrained( MODEL_FINETUNED)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_FINETUNED)
+model.eval()
+
+def filter_review(text):
+    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=128)
+    inputs = {k: v.to(device) for k, v in inputs.items()}
+    with torch.no_grad():
+        outputs = model(**inputs)
+        probs = F.softmax(outputs.logits, dim=1)
+        pred = torch.argmax(probs, dim=1).item()
+    label = "spam" if pred == 1 else "non-spam"
+    confidence = probs[0][pred].item()
+    return label, confidence

app/main.py

@@ -0,0 +1,24 @@
+from fastapi import FastAPI
+from pydantic import BaseModel
+from app.filter_review import filter_review
+import os
+
+os.environ['TRANSFORMERS_CACHE'] = '/tmp/hf_cache'
+os.makedirs('/tmp/hf_cache', exist_ok=True)
+
+app = FastAPI()
+
+@app.get("/")
+def read_root():
+    return {"message": "Selamat datang di API Filter Ulasan Spam untuk ArTour!"}
+
+class FilterReviewRequest(BaseModel):
+    text: str
+
+@app.post("/filter-review")
+def filter_spam(request: FilterReviewRequest):
+    label, confidence = filter_review(request.text)
+    
+    binary_label = 1 if label.lower() == "spam" else 0
+    
+    return {"label": binary_label, "confidence": confidence}

app/model.py

@@ -0,0 +1,8 @@
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+
+MODEL_NAME = "indobenchmark/indobert-base-p1"
+MODEL_FINETUNED = "yondikavl/artour-spam-filter"
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForSequenceClassification.from_pretrained(MODEL_FINETUNED) 
+model.eval()

requirements.txt

@@ -0,0 +1,5 @@
+fastapi
+uvicorn
+transformers
+pydantic
+https://download.pytorch.org/whl/cpu/torch-2.1.2%2Bcpu-cp310-cp310-linux_x86_64.whl