From 3e77fd44304a67a2b2253b4e56fede9762bb8464 Mon Sep 17 00:00:00 2001
From: speech_asr <wangjiaming.wjm@alibaba-inc.com>
Date: 星期四, 20 四月 2023 16:41:22 +0800
Subject: [PATCH] update
---
funasr/bin/train.py | 71 +++++++++++++++++++++++++++++++++++
1 files changed, 70 insertions(+), 1 deletions(-)
diff --git a/funasr/bin/train.py b/funasr/bin/train.py
index 94dc75c..c6f19b6 100644
--- a/funasr/bin/train.py
+++ b/funasr/bin/train.py
@@ -1,10 +1,24 @@
+import argparse
+import logging
+import os
import sys
+from io import BytesIO
import torch
+from funasr.torch_utils.model_summary import model_summary
+from funasr.torch_utils.pytorch_version import pytorch_cudnn_version
+from funasr.torch_utils.set_all_random_seed import set_all_random_seed
from funasr.utils import config_argparse
+from funasr.utils.build_args import build_args
+from funasr.utils.build_dataloader import build_dataloader
from funasr.utils.build_distributed import build_distributed
+from funasr.utils.build_model import build_model
+from funasr.utils.build_optimizer import build_optimizer
+from funasr.utils.build_scheduler import build_scheduler
+from funasr.utils.prepare_data import prepare_data
from funasr.utils.types import str2bool
+from funasr.utils.yaml_no_alias_safe_dump import yaml_no_alias_safe_dump
def get_parser():
@@ -21,6 +35,7 @@
help="The number of gpus. 0 indicates CPU mode",
)
parser.add_argument("--seed", type=int, default=0, help="Random seed")
+ parser.add_argument("--task_name", type=str, default="asr", help="Name for different tasks")
# ddp related
parser.add_argument(
@@ -259,6 +274,12 @@
action="append",
default=[],
)
+ parser.add_argument(
+ "--use_preprocessor",
+ type=str2bool,
+ default=True,
+ help="Apply preprocessing to data or not",
+ )
# pai related
parser.add_argument(
@@ -317,10 +338,58 @@
if __name__ == '__main__':
parser = get_parser()
args = parser.parse_args()
+ task_args = build_args(args)
+ args = argparse.Namespace(**vars(args), **vars(task_args))
+ # set random seed
+ set_all_random_seed(args.seed)
+ torch.backends.cudnn.enabled = args.cudnn_enabled
+ torch.backends.cudnn.benchmark = args.cudnn_benchmark
+ torch.backends.cudnn.deterministic = args.cudnn_deterministic
+
+ # ddp init
args.distributed = args.dist_world_size > 1
distributed_option = build_distributed(args)
- #
+ # for logging
+ if not distributed_option.distributed or distributed_option.dist_rank == 0:
+ logging.basicConfig(
+ level="INFO",
+ format=f"[{os.uname()[1].split('.')[0]}]"
+ f" %(asctime)s (%(module)s:%(lineno)d) %(levelname)s: %(message)s",
+ )
+ else:
+ logging.basicConfig(
+ level="ERROR",
+ format=f"[{os.uname()[1].split('.')[0]}]"
+ f" %(asctime)s (%(module)s:%(lineno)d) %(levelname)s: %(message)s",
+ )
+ # prepare files for dataloader
+ prepare_data(args, distributed_option)
+ model = build_model(args)
+ optimizer = build_optimizer(args, model=model)
+ scheduler = build_scheduler(args, optimizer)
+
+ logging.info("world size: {}, rank: {}, local_rank: {}".format(distributed_option.dist_world_size,
+ distributed_option.dist_rank,
+ distributed_option.local_rank))
+ logging.info(pytorch_cudnn_version())
+ logging.info(model_summary(model))
+ logging.info("Optimizer: {}".format(optimizer))
+ logging.info("Scheduler: {}".format(scheduler))
+
+ # dump args to config.yaml
+ if not distributed_option.distributed or distributed_option.dist_rank == 0:
+ os.makedirs(args.output_dir, exist_ok=True)
+ with open(os.path.join(args.output_dir, "config.yaml"), "w") as f:
+ logging.info("Saving the configuration in {}/{}".format(args.output_dir, "config.yaml"))
+ if args.use_pai:
+ buffer = BytesIO()
+ torch.save({"config": vars(args)}, buffer)
+ args.oss_bucket.put_object(os.path.join(args.output_dir, "config.dict"), buffer.getvalue())
+ else:
+ yaml_no_alias_safe_dump(vars(args), f, indent=4, sort_keys=False)
+
+ train_dataloader, valid_dataloader = build_dataloader(args)
--
Gitblit v1.9.1