Upload app.py with huggingface_hub

app.py

@@ -7,13 +7,19 @@ from langchain_core.output_parsers import StrOutputParser
 from langchain_huggingface import HuggingFacePipeline
 from fastapi.middleware.cors import CORSMiddleware
 import torch
+import transformers
 
+# Reduce logs
+transformers.logging.set_verbosity_error()
+
+# Create FastAPI app
 app = FastAPI(
     title="Phi-3 Code Explainer API",
     description="Explains Python code using the Phi-3 Mini model and LangChain.",
     version="1.0"
 )
 
+# CORS settings
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
@@ -22,28 +28,38 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
+# Load model/tokenizer safely for CPU-only spaces
 model_id = "microsoft/phi-3-mini-4k-instruct"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
-model = AutoModelForCausalLM.from_pretrained(
-    model_id, torch_dtype=torch.float16, device_map="auto"
-)
 
+# If GPU available use float16, else default
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model = AutoModelForCausalLM.from_pretrained(model_id).to(device)
+
+# Generation pipeline
 generator = pipeline(
     "text-generation",
     model=model,
     tokenizer=tokenizer,
     max_new_tokens=256,
     return_full_text=False,
-    do_sample=False
+    do_sample=True,
+    temperature=0.7
 )
 
+# LangChain pipeline
 llm = HuggingFacePipeline(pipeline=generator)
-
 prompt = ChatPromptTemplate.from_template(
-    "### Instruction:\nExplain this Python code step-by-step:\nUse code with caution\n{code}\n\n### Explanation:"
-)
+    "### Instruction:
+Explain this Python code step-by-step:
+Use code with caution
+{code}
 
+### Explanation:"
+)
 parser = StrOutputParser()
 chain = prompt | llm | parser
 
+# Add route
 add_routes(app, chain, path="/explain")
+