fix

Author: zigzagcai

ci_scripts/train/load_ckpt.sh

@@ -22,7 +22,7 @@ if [[ ! -f ${file} ]]; then
         exit_code=$(($exit_code + 1))
     fi
 
-srun -p ${SLURM_PARTITION} --kill-on-bad-exit=1 --exclusive --job-name=$2 -n 8 --ntasks-per-node=8 --gpus-per-task=1 python internlm/launcher/launch.py --config ${file}
+srun -p ${SLURM_PARTITION} --kill-on-bad-exit=1 --exclusive --job-name=$2 -n 8 --ntasks-per-node=8 --gpus-per-task=1 python internlm/launch/launcher.py --config ${file}
 [[ $? -ne 0 ]] && { echo "test slurm training failed.";  exit_code=$(($exit_code + 1)); }
 
 

ci_scripts/train/slurm_train.sh

@@ -22,7 +22,7 @@ if [[ -d ${CKPTS20_PATH} ]]; then
     fi
 fi
 
-srun -p ${SLURM_PARTITION} --kill-on-bad-exit=1 --exclusive --job-name=$1 -n 8 --ntasks-per-node=8 --gpus-per-task=1 python internlm/launcher/launch.py --config ./ci_scripts/train/ci_7B_sft.py
+srun -p ${SLURM_PARTITION} --kill-on-bad-exit=1 --exclusive --job-name=$1 -n 8 --ntasks-per-node=8 --gpus-per-task=1 python internlm/launch/launcher.py --config ./ci_scripts/train/ci_7B_sft.py
 [[ $? -ne 0 ]] && { echo "test slurm training failed.";  exit_code=$(($exit_code + 1)); }
 
 num=$(num_files "${CKPTS20_OUTPUT}")

ci_scripts/train/torchrun.sh

@@ -22,7 +22,7 @@ if [[ -d ${CKPTS20_PATH} ]]; then
     fi
 fi
 
-srun -p ${SLURM_PARTITION} --kill-on-bad-exit=1 --exclusive --job-name=$1 -N 1 torchrun --nnodes=1 --nproc_per_node=8 --master_port=29501 internlm/launcher/launch.py --config ./ci_scripts/train/ci_7B_sft.py --launcher torch
+srun -p ${SLURM_PARTITION} --kill-on-bad-exit=1 --exclusive --job-name=$1 -N 1 torchrun --nnodes=1 --nproc_per_node=8 --master_port=29501 internlm/launch/launcher.py --config ./ci_scripts/train/ci_7B_sft.py --launcher torch
 [[ $? -ne 0 ]] && { echo "test torch training failed.";  exit_code=$(($exit_code + 1)); }
 
 num=$(num_files "${CKPTS_OUTPUT}")

internlm/core/scheduler/base_scheduler.py

@@ -8,8 +8,8 @@
 
 import torch
 
-from internlm.inference import InferenceParams
 from internlm.core.engine import Engine
+from internlm.inference import InferenceParams
 
 
 class BaseScheduler(ABC):

internlm/inference/__init__.py

@@ -1,5 +1,5 @@
-from .inference_utils import InferenceParams, process_parallel_output
 from .inference import SequenceGenerator, batch_tokenize
+from .inference_utils import InferenceParams, process_parallel_output
 
 __all__ = [
     "InferenceParams",

internlm/inference/inference.py

@@ -4,10 +4,10 @@
 import torch
 import torch.nn.functional as F
 
-from internlm.inference import InferenceParams, process_parallel_output
 from internlm.core.context import ParallelMode  # noqa: E402
 from internlm.core.context import global_context as gpc  # noqa: E402
 from internlm.core.trainer import Trainer
+from internlm.inference import InferenceParams, process_parallel_output
 
 
 class SequenceGenerator:

internlm/launch/launcher.py

@@ -2,12 +2,12 @@
 # -*- encoding: utf-8 -*-
 
 from internlm.core.context import global_context as gpc
-from internlm.launch.trainer_builder import TrainerBuilder
 from internlm.data import (
     build_train_loader_with_data_type,
     build_valid_loader_with_data_type,
 )
 from internlm.initialize import initialize_launcher
+from internlm.launch.trainer_builder import TrainerBuilder
 from internlm.model.model_implementations.builder import create_model
 from internlm.model.model_implementations.registry import register_model_initializer
 from internlm.monitor import internevo_monitor

internlm/monitor/monitor.py

@@ -1,4 +1,3 @@
-from datetime import datetime
 import fcntl
 import logging
 import os
@@ -8,6 +7,7 @@
 import time
 import traceback
 from contextlib import contextmanager
+from datetime import datetime
 from functools import wraps
 from threading import Thread
 
@@ -16,7 +16,6 @@
 from internlm.monitor import send_feishu_msg_with_webhook
 from internlm.utils.common import SingletonMeta, set_env_var
 
-
 logger = logging.getLogger(__file__)
 internlm_accelerator = get_accelerator()