restore normal endpoint

main.py

@@ -70,7 +70,7 @@ class GenerationOutput(BaseModel):
 class ErrorOutput(BaseModel):
     error: str
 
-@app.post('/generate', response_model=Union[Response, ErrorOutput])
+@app.post('/generate-stream', response_model=Union[Response, ErrorOutput])
 def generate(request: GenerationRequest):
     t = time.perf_counter()
     print(request)
@@ -114,5 +114,39 @@ def generate(request: GenerationRequest):
             os.kill(mainpid, signal.SIGTERM)
         return {"error": str(e)}
 
+@app.post('/generate', response_model=Union[GenerationOutput, ErrorOutput])
+def generate(request: GenerationRequest):
+    t = time.perf_counter()
+    print(request)
+    try:
+        output = sanitize_input(request)
+        
+        if output[0]:
+            request = output[1]
+        else:
+            return ErrorOutput(error=output[1])
+        
+        images = model.sample(request)
+        for x in range(len(images)):
+            image = simplejpeg.encode_jpeg(images[x], quality=95) 
+            #get base64 of image
+            image = base64.b64encode(image).decode("ascii")
+            images[x] = image
+
+        process_time = time.perf_counter() - t
+        logger.info(f"Request took {process_time:0.3f} seconds")
+        return GenerationOutput(output=images)
+
+    except Exception as e:
+        capture_exception(e)
+        logger.error(str(e))
+        e_s = str(e)
+        gc.collect()
+        if "CUDA out of memory" in e_s or \
+                "an illegal memory access" in e_s or "CUDA" in e_s:
+            logger.error("GPU error, committing seppuku.")
+            os.kill(mainpid, signal.SIGTERM)
+        return {"error": str(e)}
+
 if __name__ == "__main__":
     uvicorn.run("main:app", host="0.0.0.0", port=80, log_level="info")
\ No newline at end of file