--gpu_ids 0 lose efficacy after pull

accelerate launch --mixed_precision bf16 --num_cpu_threads_per_process 1 --num_processes 1 --gpu_ids 0 sd-scripts/flux_train_network.py --pretrained_model_name_or_path "D:\pinokio\api\ajfluxlora\models\unet\bdsqlsz\flux1-dev2pro-single\flux1-dev2pro.safetensors" --clip_l "D:\pinokio\api\ajfluxlora\models\clip\clip_l.safetensors" --t5xxl "D:\pinokio\api\ajfluxlora\models\clip\t5xxl_fp16.safetensors" --ae "D:\pinokio\api\ajfluxlora\models\vae\ae.sft" --cache_latents_to_disk --save_model_as safetensors --sdpa --persistent_data_loader_workers --max_data_loader_n_workers 32 --seed 42 --gradient_checkpointing --mixed_precision bf16 --save_precision bf16 --network_module networks.lora_flux --network_dim 128 --optimizer_type adamw8bit --sample_prompts="D:\pinokio\api\ajfluxlora\outputs\dim-test\sample_prompts.txt" --sample_every_n_steps="49" --learning_rate 1e-4 --cache_text_encoder_outputs --cache_text_encoder_outputs_to_disk --fp8_base --highvram --max_train_epochs 2 --save_every_n_epochs 1 --dataset_config "D:\pinokio\api\ajfluxlora\outputs\dim-test\dataset.toml" --output_dir "D:\pinokio\api\ajfluxlora\outputs\dim-test" --output_name dim-test --timestep_sampling shift --discrete_flow_shift 3.1582 --model_prediction_type raw --guidance_scale 1 --loss_type l2
Traceback (most recent call last):
File "d:\pinokio\bin\miniconda\lib\runpy.py", line 196, in _run_module_as_main
return _run_code(code, main_globals, None,
File "d:\pinokio\bin\miniconda\lib\runpy.py", line 86, in _run_code
exec(code, run_globals)
File "D:\pinokio\api\ajfluxlora\env\Scripts\accelerate.exe\__main__.py", line 7, in
File "d:\pinokio\api\ajfluxlora\env\lib\site-packages\accelerate\commands\accelerate_cli.py", line 48, in main
args.func(args)
File "d:\pinokio\api\ajfluxlora\env\lib\site-packages\accelerate\commands\launch.py", line 1084, in launch_command
args, defaults, mp_from_config_flag = _validate_launch_command(args)
File "d:\pinokio\api\ajfluxlora\env\lib\site-packages\accelerate\commands\launch.py", line 957, in _validate_launch_command
raise ValueError(
ValueError: Less than two GPU ids were configured and tried to run on on multiple GPUs. Please ensure at least two are specified for `--gpu_ids`, or use `--gpu_ids='all'`.

This command was usable before, but after the update, it is not possible to specify a specific card for training

Reproduce the command using accelerate launch and sd-scripts/flux_train_network.py, then compare the launch behavior before and after the reported update. Trace how --gpu_ids 0 is validated and passed to the training entry point. Done means selecting that single GPU no longer raises the multiple-GPU validation error.

Written by the indexing model from the issue text.