"FAILED: cpu_adam.so " when I use LoRA to train the model😭😭😭

When I run

deepspeed fastchat/train/train_lora.py     --model_name_or_path /root/autodl-tmp/cjk/Fast-Chat-main/Codellama-7B     --lora_r 16     --lora_alpha 16     --lora_dropout 0.05     --data_path /root/autodl-tmp/cjk/Fast-Chat-main/data/Tool_ReAct_train_bird_Qshot.json     --bf16 True     --output_dir ./checkpoints     --num_train_epochs 8     --per_device_train_batch_size 1     --per_device_eval_batch_size 1     --gradient_accumulation_steps 1     --evaluation_strategy "no"     --save_strategy "steps"     --save_steps 1200     --save_total_limit 100     --learning_rate 2e-4     --weight_decay 0.     --warmup_ratio 0.03     --lr_scheduler_type "cosine"     --logging_steps 1     --tf32 True     --model_max_length 4096     --q_lora True     --deepspeed playground/deepspeed_config_s2.json

I got the error below:

FAILED: cpu_adam.so 
c++ cpu_adam.o cpu_adam_impl.o -shared -lcurand -L/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/torch/lib -lc10 -ltorch_cpu -ltorch -ltorch_python -o cpu_adam.so
/usr/bin/ld: cannot find -lcurand: No such file or directory
collect2: error: ld returned 1 exit status
ninja: build stopped: subcommand failed.
Traceback (most recent call last):
  File "/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 2107, in _run_ninja_build
    subprocess.run(
  File "/root/miniconda3/envs/fastchat/lib/python3.10/subprocess.py", line 524, in run
    raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1.

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/root/autodl-tmp/cjk/Fast-Chat-main/fastchat/train/train_lora.py", line 44, in <module>
    deepspeed.ops.op_builder.CPUAdamBuilder().load()
  File "/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/deepspeed/ops/op_builder/builder.py", line 508, in load
    return self.jit_load(verbose)
  File "/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/deepspeed/ops/op_builder/builder.py", line 555, in jit_load
    op_module = load(name=self.name,
  File "/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 1309, in load
    return _jit_compile(
  File "/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 1719, in _jit_compile
    _write_ninja_file_and_build_library(
  File "/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 1832, in _write_ninja_file_and_build_library
    _run_ninja_build(
  File "/root/miniconda3/envs/fastchat/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 2123, in _run_ninja_build
    raise RuntimeError(message) from e
RuntimeError: Error building extension 'cpu_adam'

Here are some information on my device:

PyTorch  2.1.0
Python  3.10(ubuntu22.04)
Cuda  12.1

And I was strictly following the step in README, and those packages were all installed sucessfully, who can help me😭😭😭

Start at fastchat/train/train_lora.py, especially the deepspeed.ops.op_builder.CPUAdamBuilder().load() call, and reproduce the command using playground/deepspeed_config_s2.json. Inspect the reported linker failure for the missing libcurand dependency and the stated PyTorch, Python, CUDA, and Ubuntu versions; done means the CPUAdam extension builds and LoRA training starts successfully.