fix server config

Author: tastelikefeet

cookbook/client/server/megatron/server_config.yaml

@@ -42,7 +42,7 @@ applications:
     import_path: sampler
     args:
       model_id: "ms://Qwen/Qwen3.5-27B"  # ModelScope model identifier
-      nproc_per_node: 8               # Number of GPU processes per node
+      nproc_per_node: 4               # Number of GPU processes per node
       sampler_type: vllm              # Inference engine: 'vllm' (fast) or 'torch' (TorchSampler)
       engine_args:                    # vLLM engine-specific settings
         max_model_len: 32000           # Maximum sequence length the engine supports
@@ -84,7 +84,7 @@ applications:
       model_id: "ms://Qwen/Qwen3.5-27B" # ModelScope model identifier
       max_length: 32000                           # model max length
       max_loras: 5                                # model max loras
-      nproc_per_node: 8                           # Number of GPU processes per node
+      nproc_per_node: 4                           # Number of GPU processes per node
       device_group:
         name: model
         ranks: 4       # GPU rank indices

pyproject.toml

@@ -10,7 +10,7 @@ dependencies = [
   "datasets>=3.0,<4.0",
   "omegaconf>=2.3.0,<3.0.0",
   "fastapi",
-  "modelscope[framework]>=1.34.0",
+  "modelscope[framework]>=1.35.0",
   "safetensors",
   "peft>=0.11.0,<=0.19.0",
   "transformers",