423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-17 12:20:37 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix #4137

Browse Source
This commit is contained in:
hiyouga
2024-06-07 19:16:06 +08:00
parent cce0fad91c
commit 8bf9da659c
2 changed files with 5 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/llamafactory/cli.py
View File

@@ -72,12 +72,8 @@ def main():
elif command == Command.EXPORT:
export_model()
elif command == Command.TRAIN:
disable_torchrun = os.environ.get("TORCHRUN_DISABLED", "0").lower() in ["true", "1"]
if disable_torchrun and get_device_count() > 1:
logger.warning("`torchrun` cannot be disabled when device count > 1.")
disable_torchrun = False
if (not disable_torchrun) and (get_device_count() > 0):
force_torchrun = os.environ.get("FORCE_TORCHRUN", "0").lower() in ["true", "1"]
if force_torchrun or get_device_count() > 1:
master_addr = os.environ.get("MASTER_ADDR", "127.0.0.1")
master_port = os.environ.get("MASTER_PORT", str(random.randint(20001, 29999)))
logger.info("Initializing distributed tasks at: {}:{}".format(master_addr, master_port))