From ec5948bbb7ab9fd6fd00820be24f4315bb69263d Mon Sep 17 00:00:00 2001 From: Jintao Huang Date: Thu, 26 Dec 2024 11:09:01 +0800 Subject: [PATCH] update --- examples/deploy/client/README.md | 1 - examples/deploy/client/llm/swift_client.py | 1 + examples/deploy/client/mllm/swift_client.py | 1 + 3 files changed, 2 insertions(+), 1 deletion(-) delete mode 100644 examples/deploy/client/README.md diff --git a/examples/deploy/client/README.md b/examples/deploy/client/README.md deleted file mode 100644 index 5198b501b..000000000 --- a/examples/deploy/client/README.md +++ /dev/null @@ -1 +0,0 @@ -In each client `.py` program, we have added the `run_deploy` context. The `run_deploy` function is a simple way to deploy locally, making it convenient for users to run the program directly. In common deployment scenarios, users only need to remove the deployment context and modify the `host` and `port` in order to use the client. diff --git a/examples/deploy/client/llm/swift_client.py b/examples/deploy/client/llm/swift_client.py index bad18dd86..16df8d3af 100644 --- a/examples/deploy/client/llm/swift_client.py +++ b/examples/deploy/client/llm/swift_client.py @@ -53,6 +53,7 @@ def run_client(host: str = '127.0.0.1', port: int = 8000): DeployArguments) from swift.plugin import InferStats # TODO: The current 'pt' deployment does not support automatic batch. + # NOTE: In a real deployment scenario, please comment out the context of run_deploy. with run_deploy( DeployArguments(model='Qwen/Qwen2.5-1.5B-Instruct', verbose=False, log_interval=-1, infer_backend='vllm')) as port: diff --git a/examples/deploy/client/mllm/swift_client.py b/examples/deploy/client/mllm/swift_client.py index 2dc76e689..85f7f418c 100644 --- a/examples/deploy/client/mllm/swift_client.py +++ b/examples/deploy/client/mllm/swift_client.py @@ -118,6 +118,7 @@ def run_client(host: str = '127.0.0.1', port: int = 8000): DeployArguments) from swift.plugin import InferStats # TODO: The current 'pt' deployment does not support automatic batch. + # NOTE: In a real deployment scenario, please comment out the context of run_deploy. with run_deploy( DeployArguments(model='Qwen/Qwen2-VL-2B-Instruct', verbose=False, log_interval=-1, infer_backend='vllm')) as port: