Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -11,6 +11,7 @@ load_dotenv()
|
|
11 |
|
12 |
app = FastAPI()
|
13 |
|
|
|
14 |
global_data = {
|
15 |
'models': {},
|
16 |
'tokens': {
|
@@ -75,7 +76,7 @@ class ModelManager:
|
|
75 |
self.loaded = True
|
76 |
return models
|
77 |
except Exception as e:
|
78 |
-
print(f"Error loading
|
79 |
return []
|
80 |
|
81 |
model_manager = ModelManager()
|
@@ -113,49 +114,34 @@ def remove_repetitive_responses(responses):
|
|
113 |
unique_responses.append(response)
|
114 |
return unique_responses
|
115 |
|
116 |
-
def generate_chat_response(request, model_data):
|
117 |
-
model = model_data['model']
|
118 |
-
try:
|
119 |
-
user_input = normalize_input(request.message)
|
120 |
-
response = model(user_input, top_k=request.top_k, top_p=request.top_p, temperature=request.temperature)
|
121 |
-
return response
|
122 |
-
except Exception as e:
|
123 |
-
print(f"Error generating response with model {model_data['name']}: {e}")
|
124 |
-
return None
|
125 |
-
|
126 |
-
@app.post("/generate")
|
127 |
@spaces.GPU(duration=0)
|
|
|
128 |
async def generate(request: ChatRequest):
|
129 |
try:
|
130 |
-
|
131 |
-
|
132 |
-
for model_data in models:
|
133 |
-
response = generate_chat_response(request, model_data)
|
134 |
-
if response:
|
135 |
-
responses.append({
|
136 |
-
"model": model_data['name'],
|
137 |
-
"response": response
|
138 |
-
})
|
139 |
-
|
140 |
-
if not responses:
|
141 |
-
raise HTTPException(status_code=500, detail="Error: No responses generated.")
|
142 |
|
143 |
-
|
|
|
|
|
144 |
best_response = responses[0] if responses else {}
|
|
|
145 |
return {
|
146 |
"best_response": best_response,
|
147 |
-
"all_responses":
|
148 |
}
|
149 |
-
except Exception:
|
150 |
-
|
|
|
151 |
|
152 |
@app.api_route("/{method_name:path}", methods=["GET", "POST", "PUT", "DELETE", "PATCH"])
|
153 |
async def handle_request(method_name: str, request: Request):
|
154 |
try:
|
155 |
body = await request.json()
|
156 |
return {"message": "Request handled successfully", "body": body}
|
157 |
-
except Exception:
|
158 |
-
|
|
|
159 |
|
160 |
if __name__ == "__main__":
|
161 |
uvicorn.run(app, host="0.0.0.0", port=7860)
|
|
|
11 |
|
12 |
app = FastAPI()
|
13 |
|
14 |
+
# Global data storage
|
15 |
global_data = {
|
16 |
'models': {},
|
17 |
'tokens': {
|
|
|
76 |
self.loaded = True
|
77 |
return models
|
78 |
except Exception as e:
|
79 |
+
print(f"Error loading models: {e}")
|
80 |
return []
|
81 |
|
82 |
model_manager = ModelManager()
|
|
|
114 |
unique_responses.append(response)
|
115 |
return unique_responses
|
116 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
117 |
@spaces.GPU(duration=0)
|
118 |
+
@app.post("/generate")
|
119 |
async def generate(request: ChatRequest):
|
120 |
try:
|
121 |
+
if not global_data['models']:
|
122 |
+
raise HTTPException(status_code=500, detail="Models not loaded")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
123 |
|
124 |
+
model = global_data['models'][0]['model']
|
125 |
+
inputs = normalize_input(request.message)
|
126 |
+
responses = model.generate(inputs, top_k=request.top_k, top_p=request.top_p, temperature=request.temperature)
|
127 |
best_response = responses[0] if responses else {}
|
128 |
+
unique_responses = remove_repetitive_responses(responses)
|
129 |
return {
|
130 |
"best_response": best_response,
|
131 |
+
"all_responses": unique_responses
|
132 |
}
|
133 |
+
except Exception as e:
|
134 |
+
print(f"Error in generate endpoint: {e}")
|
135 |
+
return {"error": str(e)}
|
136 |
|
137 |
@app.api_route("/{method_name:path}", methods=["GET", "POST", "PUT", "DELETE", "PATCH"])
|
138 |
async def handle_request(method_name: str, request: Request):
|
139 |
try:
|
140 |
body = await request.json()
|
141 |
return {"message": "Request handled successfully", "body": body}
|
142 |
+
except Exception as e:
|
143 |
+
print(f"Error handling request: {e}")
|
144 |
+
return {"error": str(e)}
|
145 |
|
146 |
if __name__ == "__main__":
|
147 |
uvicorn.run(app, host="0.0.0.0", port=7860)
|