[BUG] Single node Autotune failed : Hostname '{hostname}' not found in hostfile

[oushu1zhangxiangxuan1]

Describe the bug
Use deepspeed autotune error:

nohup: ignoring input
[2024-01-23 14:15:43,187] [INFO] [real_accelerator.py:158:get_accelerator] Setting ds_accelerator to cuda (auto detect)
/root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: /root/space/conda_envs/hubc12 did not contain ['libcudart.so', 'libcudart.so.11.0', 'libcudart.so.12.0'] as expected! Searching further paths...
  warn(msg)
/root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: WARNING: Compute capability < 7.5 detected! Only slow 8-bit matmul is supported for your GPU!
  warn(msg)

===================================BUG REPORT===================================
Welcome to bitsandbytes. For bug reports, please run

python -m bitsandbytes

 and submit this information together with your error trace to: https://github.com/TimDettmers/bitsandbytes/issues
================================================================================
bin /root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/libbitsandbytes_cuda121_nocublaslt.so
CUDA_SETUP: WARNING! libcudart.so not found in any environmental path. Searching in backup paths...
CUDA SETUP: CUDA runtime path found: /usr/local/cuda/lib64/libcudart.so
CUDA SETUP: Highest compute capability among GPUs detected: 7.0
CUDA SETUP: Detected CUDA version 121
CUDA SETUP: Loading binary /root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/libbitsandbytes_cuda121_nocublaslt.so...
[2024-01-23 14:15:46,018] [INFO] [runner.py:464:main] Using IP address of  for node v1004
[2024-01-23 14:15:46,019] [INFO] [autotuner.py:71:__init__] Created autotuning experiments directory: autotuning_exps
[2024-01-23 14:15:46,019] [INFO] [autotuner.py:84:__init__] Created autotuning results directory: autotuning_exps
[2024-01-23 14:15:46,019] [INFO] [autotuner.py:200:_get_resource_manager] active_resources = OrderedDict([('v1004', [0, 1, 2, 3])])
[2024-01-23 14:15:46,019] [INFO] [runner.py:363:run_autotuning] [Start] Running autotuning
[2024-01-23 14:15:46,020] [INFO] [autotuner.py:413:tune] Fast mode is enabled. Tuning micro batch size only.
[2024-01-23 14:15:46,020] [INFO] [autotuner.py:669:model_info_profile_run] Starting model info profile run.

  0%|          | 0/1 [00:00<?, ?it/s][2024-01-23 14:15:46,037] [INFO] [scheduler.py:344:run_experiment] Scheduler wrote ds_config to autotuning_results/profile_model_info/ds_config.json, /mnt/datadisk0/repos/eosphoros-ai/DB-GPT-Hub/autotuning_results/profile_model_info/ds_config.json
[2024-01-23 14:15:46,040] [INFO] [scheduler.py:351:run_experiment] Scheduler wrote exp to autotuning_results/profile_model_info/exp.json, /mnt/datadisk0/repos/eosphoros-ai/DB-GPT-Hub/autotuning_results/profile_model_info/exp.json
[2024-01-23 14:15:46,043] [INFO] [scheduler.py:378:run_experiment] Launching exp_id = 0, exp_name = profile_model_info, with resource = v1004:0,1,2,3, and ds_config = /mnt/datadisk0/repos/eosphoros-ai/DB-GPT-Hub/autotuning_results/profile_model_info/ds_config.json
[2024-01-23 14:15:51,032] [INFO] [scheduler.py:430:clean_up] Done cleaning up exp_id = 0 on the following workers: v1004
[2024-01-23 14:15:51,032] [INFO] [scheduler.py:393:run_experiment] Done running exp_id = 0, exp_name = profile_model_info, with resource = v1004:0,1,2,3

100%|██████████| 1/1 [00:05<00:00,  5.00s/it]
100%|██████████| 1/1 [00:10<00:00, 10.03s/it]
[2024-01-23 14:15:56,058] [ERROR] [autotuner.py:699:model_info_profile_run] The model is not runnable with DeepSpeed with error = (f"Hostname '{hostname}' not found in hostfile")

[2024-01-23 14:15:56,058] [INFO] [runner.py:368:run_autotuning] [End] Running autotuning

To Reproduce
Run cmd

deepspeed  --autotuning=tune --hostfile=/xxxxx/hostfile --include=v1004:0,1,2,3  xxx/sft_train.py \
    --deepspeed /root/space/xxx/configs/ds_config_auto.json \
    --model_name_or_path /root/space/models/codellama/CodeLlama-13b-Instruct-hf/main \
    --do_train \
    --dataset $traindata \
    --max_source_length 4096 \
    --max_target_length 256 \
    --template llama2 \
    --finetuning_type lora \
    --lora_rank 2 \
    --lora_alpha 32 \
    --lora_target q_proj \
    --output_dir dbgpt_hub/output/adapter/CodeLlama-13b-Instruct-hf-$(date +"%Y%m%d_%H%M%S")-xds-auto-qv-${traindata}/ \
    --overwrite_cache \
    --overwrite_output_dir \
    --per_device_train_batch_size 1 \
    --gradient_accumulation_steps 1 \
    --lr_scheduler_type cosine_with_restarts \
    --logging_steps 25 \
    --save_steps 1 \
    --learning_rate 2e-4 \
    --num_train_epochs 10000 \
    --plot_loss

df_config

{
    "train_micro_batch_size_per_gpu": "auto",
    "fp16": {
      "enabled": true
    },
    "autotuning": {
      "enabled": true,
      "arg_mappings": {
        "train_micro_batch_size_per_gpu": "--per_device_train_batch_size",
        "gradient_accumulation_steps ": "--gradient_accumulation_steps"
      }
    }
  }

ds_report output

[2024-01-23 17:48:25,600] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)

===================================BUG REPORT===================================
Welcome to bitsandbytes. For bug reports, please run

python -m bitsandbytes

 and submit this information together with your error trace to: https://github.com/TimDettmers/bitsandbytes/issues
================================================================================
bin /root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/libbitsandbytes_cuda121_nocublaslt.so
/root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: /root/space/conda_envs/hubc12 did not contain ['libcudart.so', 'libcudart.so.11.0', 'libcudart.so.12.0'] as expected! Searching further paths...
  warn(msg)
CUDA_SETUP: WARNING! libcudart.so not found in any environmental path. Searching in backup paths...
CUDA SETUP: CUDA runtime path found: /usr/local/cuda/lib64/libcudart.so
CUDA SETUP: Highest compute capability among GPUs detected: 7.0
CUDA SETUP: Detected CUDA version 121
/root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: WARNING: Compute capability < 7.5 detected! Only slow 8-bit matmul is supported for your GPU!
  warn(msg)
CUDA SETUP: Loading binary /root/space/conda_envs/hubc12/lib/python3.10/site-packages/bitsandbytes/libbitsandbytes_cuda121_nocublaslt.so...
--------------------------------------------------
DeepSpeed C++/CUDA extension op report
--------------------------------------------------
NOTE: Ops not installed will be just-in-time (JIT) compiled at
      runtime if needed. Op compatibility means that your system
      meet the required dependencies to JIT install the op.
--------------------------------------------------
JIT compiled ops requires ninja
ninja .................. [OKAY]
--------------------------------------------------
op name ................ installed .. compatible
--------------------------------------------------
 [WARNING]  async_io requires the dev libaio .so object and headers but these were not found.
 [WARNING]  async_io: please install the libaio-devel package with yum
 [WARNING]  If libaio is already installed (perhaps from source), try setting the CFLAGS and LDFLAGS environment variables to where it can be found.
async_io ............... [NO] ....... [NO]
fused_adam ............. [NO] ....... [OKAY]
cpu_adam ............... [NO] ....... [OKAY]
cpu_adagrad ............ [NO] ....... [OKAY]
cpu_lion ............... [NO] ....... [OKAY]
 [WARNING]  Please specify the CUTLASS repo directory as environment variable $CUTLASS_PATH
evoformer_attn ......... [NO] ....... [NO]
fused_lamb ............. [NO] ....... [OKAY]
fused_lion ............. [NO] ....... [OKAY]
inference_core_ops ..... [NO] ....... [OKAY]
cutlass_ops ............ [NO] ....... [OKAY]
quantizer .............. [NO] ....... [OKAY]
ragged_device_ops ...... [NO] ....... [OKAY]
ragged_ops ............. [NO] ....... [OKAY]
random_ltd ............. [NO] ....... [OKAY]
 [WARNING]  sparse_attn requires a torch version >= 1.5 and < 2.0 but detected 2.1
 [WARNING]  using untested triton version (2.1.0), only 1.0.0 is known to be compatible
sparse_attn ............ [NO] ....... [NO]
spatial_inference ...... [NO] ....... [OKAY]
transformer ............ [NO] ....... [OKAY]
stochastic_transformer . [NO] ....... [OKAY]
transformer_inference .. [NO] ....... [OKAY]
--------------------------------------------------
DeepSpeed general environment info:
torch install path ............... ['/root/space/conda_envs/hubc12/lib/python3.10/site-packages/torch']
torch version .................... 2.1.1
deepspeed install path ........... ['/root/space/conda_envs/hubc12/lib/python3.10/site-packages/deepspeed']
deepspeed info ................... 0.13.0, unknown, unknown
torch cuda version ............... 12.1
torch hip version ................ None
nvcc version ..................... 12.1
deepspeed wheel compiled w. ...... torch 2.1, cuda 12.1
shared memory (/dev/shm) size .... 78.33 GB

Screenshots

System info (please complete the following information):

• OS: centos 7
• GPU count and types 2V100 + 4V100
• Interconnects (if applicable) [e.g., two machines connected with 100 Gbps IB]
• Python version 3.10.13
• Any other relevant info about your setup

Launcher context
Are you launching your experiment with the deepspeed launcher, MPI, or something else?

Docker context
Are you using a specific docker image that you can share?

Additional context
Add any other context about the problem here.

[mrwyattii]

#4996 is merged, this problem is resolved