[CQ]

servers/blip2.py

@@ -9,7 +9,7 @@
 
 
 def get_engine_name(rank):
-    return "rank{}.engine".format(rank)
+    return f"rank{rank}.engine"
 
 
 def trt_dtype_to_torch(dtype):
@@ -20,7 +20,7 @@ def trt_dtype_to_torch(dtype):
     elif dtype == trt.int32:
         return torch.int32
     else:
-        raise TypeError("%s is not supported" % dtype)
+        raise TypeError(f"{dtype} is not supported")
 
 
 def TRTOPT(args, config):

servers/cogvlm.py

@@ -220,7 +220,7 @@ async def predict(model_id: str, params: dict):
     chunk = ChatCompletionResponse(
         model=model_id, choices=[choice_data], object="chat.completion.chunk"
     )
-    yield "{}".format(chunk.model_dump_json(exclude_unset=True))
+    yield f"{chunk.model_dump_json(exclude_unset=True)}"
 
     previous_text = ""
     for new_response in generate_stream_cogvlm(model, tokenizer, params):
@@ -238,15 +238,15 @@ async def predict(model_id: str, params: dict):
         chunk = ChatCompletionResponse(
             model=model_id, choices=[choice_data], object="chat.completion.chunk"
         )
-        yield "{}".format(chunk.model_dump_json(exclude_unset=True))
+        yield f"{chunk.model_dump_json(exclude_unset=True)}"
     choice_data = ChatCompletionResponseStreamChoice(
         index=0,
         delta=DeltaMessage(),
     )
     chunk = ChatCompletionResponse(
         model=model_id, choices=[choice_data], object="chat.completion.chunk"
     )
-    yield "{}".format(chunk.model_dump_json(exclude_unset=True))
+    yield f"{chunk.model_dump_json(exclude_unset=True)}"
 
 
 def generate_cogvlm(
@@ -405,9 +405,7 @@ def generate_stream_cogvlm(
         torch_type = torch.float16
 
     print(
-        "========Use torch type as:{} with device:{}========\n\n".format(
-            torch_type, DEVICE
-        )
+        f"========Use torch type as:{torch_type} with device:{DEVICE}========\n\n"
     )
 
     if "cuda" in DEVICE:

servers/qwen_tensort.py

@@ -30,8 +30,8 @@
 
 def get_engine_name(model, dtype, tp_size, pp_size, rank):
     if pp_size == 1:
-        return "{}_{}_tp{}_rank{}.engine".format(model, dtype, tp_size, rank)
-    return "{}_{}_tp{}_pp{}_rank{}.engine".format(model, dtype, tp_size, pp_size, rank)
+        return f"{model}_{dtype}_tp{tp_size}_rank{rank}.engine"
+    return f"{model}_{dtype}_tp{tp_size}_pp{pp_size}_rank{rank}.engine"
 
 
 def trt_dtype_to_torch(dtype):
@@ -42,7 +42,7 @@ def trt_dtype_to_torch(dtype):
     elif dtype == trt.int32:
         return torch.int32
     else:
-        raise TypeError("%s is not supported" % dtype)
+        raise TypeError(f"{dtype} is not supported")
 
 
 class QWenInfer(object):

swarms_cloud/sky_api.py

@@ -95,7 +95,7 @@ def execute(self, task: Task = None, cluster_name: str = None, **kwargs):
             _type_: _description_
         """
         if cluster_name not in self.clusters:
-            raise ValueError("Cluster {} does not exist".format(cluster_name))
+            raise ValueError(f"Cluster {cluster_name} does not exist")
         try:
             return sky.exec(
                 task=task,