fix /tokens docs && tag v1.0.1 (#753)

Co-authored-by: shihaobai <baishihao@sensetime.com>

Author: shihaobai

Dockerfile

@@ -1,4 +1,4 @@
-FROM nvidia/cuda:12.4.0-runtime-ubuntu20.04 as base
+FROM nvcr.io/nvidia/tritonserver:24.04-py3-min as base
 ARG PYTORCH_VERSION=2.5.1
 ARG PYTHON_VERSION=3.9
 ARG CUDA_VERSION=12.4
@@ -38,5 +38,7 @@ WORKDIR /root
 COPY ./requirements.txt /lightllm/requirements.txt
 RUN pip install -r /lightllm/requirements.txt --no-cache-dir --ignore-installed --extra-index-url https://download.pytorch.org/whl/cu124
 
+RUN pip install --no-cache-dir nvidia-nccl-cu12==2.25.1  # for allreduce hang issues in multinode H100
+
 COPY . /lightllm
 RUN pip install -e /lightllm --no-cache-dir

docs/CN/source/getting_started/quickstart.rst

@@ -68,6 +68,7 @@
 双机H100部署 DeepSeek-R1 模型，启动命令如下：
 
 .. code-block:: console
+
     $ # Node 0
     $ LOADWORKER=8 python -m lightllm.server.api_server --model_dir ~/models/DeepSeek-R1 --tp 16 --graph_max_batch_size 100 --nccl_host master_addr --nnodes 2 --node_rank 0
     $ # Node 1

docs/EN/source/getting_started/quickstart.rst

@@ -65,6 +65,7 @@ For the DeepSeek-R1 model on single H200, it can be launched with the following
 For the DeepSeek-R1 model on two H100, it can be launched with the following command:
 
 .. code-block:: console
+
     $ # Node 0
     $ LOADWORKER=8 python -m lightllm.server.api_server --model_dir ~/models/DeepSeek-R1 --tp 16 --graph_max_batch_size 100 --nccl_host master_addr --nnodes 2 --node_rank 0
     $ # Node 1

lightllm/server/api_http.py

@@ -326,7 +326,7 @@ async def tokens(request: Request):
     try:
         request_dict = await request.json()
         prompt = request_dict.pop("text")
-        parameters = request_dict.pop("parameters")
+        parameters = request_dict.pop("parameters", {})
         return JSONResponse({"ntokens": g_objs.httpserver_manager.tokens(prompt, parameters)}, status_code=200)
     except Exception as e:
         return create_error_response(HTTPStatus.EXPECTATION_FAILED, f"error: {str(e)}")

setup.py

@@ -3,7 +3,7 @@
 package_data = {"lightllm": ["common/all_kernel_configs/*/*.json"]}
 setup(
     name="lightllm",
-    version="1.0.0",
+    version="1.0.1",
     packages=find_packages(exclude=("build", "include", "test", "dist", "docs", "benchmarks", "lightllm.egg-info")),
     author="model toolchain",
     author_email="",