Add app, model loader, requirements.txt

app.py

@@ -0,0 +1,39 @@
+import streamlit as st
+import torch
+from PIL import Image
+from transformers import AutoFeatureExtractor, AutoModelForImageClassification
+
+# 모델 및 설정 로드
+@st.cache_resource
+def load_model():
+    feature_extractor = AutoFeatureExtractor.from_pretrained("xinyu1205/recognize-anything-plus-model")
+    model = AutoModelForImageClassification.from_pretrained("xinyu1205/recognize-anything-plus-model")
+    model.eval()
+    return feature_extractor, model
+
+# 예측 함수
+def predict(image, feature_extractor, model):
+    inputs = feature_extractor(images=image, return_tensors="pt")
+    with torch.no_grad():
+        outputs = model(**inputs)
+    
+    logits = outputs.logits
+    # 상위 5개 태그 반환
+    top_5 = torch.topk(logits, k=5)
+    return [model.config.id2label[i.item()] for i in top_5.indices[0]]
+
+# Streamlit 앱
+st.title("RAM++ Image Tagging")
+
+feature_extractor, model = load_model()
+
+uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "png", "jpeg"])
+
+if uploaded_file is not None:
+    image = Image.open(uploaded_file)
+    st.image(image, caption='Uploaded Image', use_column_width=True)
+
+    if st.button('Get Tags'):
+        tags = predict(image, feature_extractor, model)
+        st.write("Predicted Tags:")
+        st.write(", ".join(tags))

ram_plus_model.py

@@ -0,0 +1,24 @@
+from transformers import AutoFeatureExtractor, AutoModelForImageClassification
+from PIL import Image
+import torch
+
+class RAMPlusModel:
+    def __init__(self):
+        self.feature_extractor = AutoFeatureExtractor.from_pretrained("xinyu1205/recognize-anything-plus-model")
+        self.model = AutoModelForImageClassification.from_pretrained("xinyu1205/recognize-anything-plus-model")
+        self.model.eval()
+
+    def predict(self, image):
+        inputs = self.feature_extractor(images=image, return_tensors="pt")
+        with torch.no_grad():
+            outputs = self.model(**inputs)
+        
+        logits = outputs.logits
+        predicted_classes = logits.argmax(-1)
+        
+        # 상위 5개 태그 반환 (이 부분은 모델의 실제 출력에 따라 조정 필요)
+        top_5 = torch.topk(logits, k=5)
+        return [self.model.config.id2label[i.item()] for i in top_5.indices[0]]
+
+# 모델 인스턴스 생성
+model = RAMPlusModel()

requirements.txt

@@ -0,0 +1,4 @@
+streamlit
+torch
+transformers
+Pillow