Upload 2 files

app.py

@@ -0,0 +1,77 @@
+import json
+import re
+import torch
+import gradio as gr
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+MODEL_ID = "Amey9766/llama32-hotel-review-triage"
+
+def load_model():
+    tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
+    model = AutoModelForCausalLM.from_pretrained(
+        MODEL_ID,
+        device_map="auto",
+        torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32
+    )
+    model.eval()
+    return tokenizer, model
+
+tokenizer, model = load_model()
+
+def extract_json(text):
+    match = re.search(r"\{.*\}", text, re.DOTALL)
+    return match.group(0) if match else text
+
+def triage_review(review, max_tokens, temperature):
+    if not review.strip():
+        return "Please enter a hotel review."
+
+    system_prompt = "You are a hospitality review triage assistant. Output ONLY valid JSON."
+
+    messages = [
+        {"role": "system", "content": system_prompt},
+        {"role": "user", "content": review.strip()}
+    ]
+
+    prompt = tokenizer.apply_chat_template(messages, tokenize=False)
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+
+    with torch.no_grad():
+        output = model.generate(
+            **inputs,
+            max_new_tokens=max_tokens,
+            do_sample=temperature > 0,
+            temperature=temperature if temperature > 0 else None,
+            pad_token_id=tokenizer.eos_token_id
+        )
+
+    decoded = tokenizer.decode(output[0], skip_special_tokens=True)
+    return extract_json(decoded)
+
+with gr.Blocks(title="Hotel Review Triage") as demo:
+    gr.Markdown("## 🏨 Hotel Review Triage Demo")
+    gr.Markdown(
+        "Paste a hotel review below. The model will return **structured JSON** "
+        "for operational triage (department, severity, sentiment, summary)."
+    )
+
+    review_input = gr.Textbox(
+        label="Hotel Review",
+        placeholder="The room was dirty and the AC didn’t work.",
+        lines=5
+    )
+
+    with gr.Row():
+        max_tokens = gr.Slider(64, 512, value=256, step=32, label="Max new tokens")
+        temperature = gr.Slider(0.0, 1.0, value=0.0, step=0.1, label="Temperature")
+
+    output = gr.Code(label="Model Output (JSON)", language="json")
+
+    run_btn = gr.Button("Generate JSON", variant="primary")
+    run_btn.click(
+        fn=triage_review,
+        inputs=[review_input, max_tokens, temperature],
+        outputs=output
+    )
+
+demo.launch()

requirements.txt

@@ -0,0 +1,5 @@
+transformers>=4.41.0
+torch
+accelerate
+sentencepiece
+gradio