新版本打开webui的条件

[shadow01a]

之前我使用的llama-factory仓库commit为d8a5571，webui可以在无gpu时打开。这样做方便了我提前进行调整，但当我将仓库更新到最新49436e9并重新安装之后，webui似乎无法在无gpu的情况下打开了。

请问是什么更改造成了该问题？如何解决呢？以后的更新中webui可以回到原本的打开条件吗？

[shadow01a]

新版本报错如下

/root/miniconda3/envs/1/lib/python3.10/site-packages/torch/cuda/__init__.py:734: UserWarning: Can't initialize NVML
  warnings.warn("Can't initialize NVML")
The installed version of bitsandbytes was compiled without GPU support. 8-bit optimizers, 8-bit multiplication, and GPU quantization are unavailable.
Traceback (most recent call last):
  File "/root/miniconda3/envs/1/bin/llamafactory-cli", line 5, in <module>
    from llamafactory.cli import main
  File "/root/LLaMA-Factory/src/llamafactory/cli.py", line 21, in <module>
    from . import launcher
  File "/root/LLaMA-Factory/src/llamafactory/launcher.py", line 15, in <module>
    from llamafactory.train.tuner import run_exp  # use absolute import
  File "/root/LLaMA-Factory/src/llamafactory/train/tuner.py", line 31, in <module>
    from .dpo import run_dpo
  File "/root/LLaMA-Factory/src/llamafactory/train/dpo/__init__.py", line 15, in <module>
    from .workflow import run_dpo
  File "/root/LLaMA-Factory/src/llamafactory/train/dpo/workflow.py", line 26, in <module>
    from ..trainer_utils import create_modelcard_and_push, create_ref_model
  File "/root/LLaMA-Factory/src/llamafactory/train/trainer_utils.py", line 43, in <module>
    from galore_torch import GaLoreAdafactor, GaLoreAdamW, GaLoreAdamW8bit  # type: ignore
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/galore_torch/__init__.py", line 3, in <module>
    from .adamw8bit import AdamW8bit as GaLoreAdamW8bit
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/galore_torch/adamw8bit.py", line 1, in <module>
    from bitsandbytes.optim.optimizer import Optimizer2State
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/bitsandbytes/__init__.py", line 15, in <module>
    from .nn import modules
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/bitsandbytes/nn/__init__.py", line 17, in <module>
    from .triton_based_modules import (
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/bitsandbytes/nn/triton_based_modules.py", line 6, in <module>
    from bitsandbytes.triton.dequantize_rowwise import dequantize_rowwise
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/bitsandbytes/triton/dequantize_rowwise.py", line 36, in <module>
    def _dequantize_rowwise(
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 368, in decorator
    return Autotuner(fn, fn.arg_names, configs, key, reset_to_zero, restore_value, pre_hook=pre_hook,
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 130, in __init__
    self.do_bench = driver.active.get_benchmarker()
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/triton/runtime/driver.py", line 23, in __getattr__
    self._initialize_obj()
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/triton/runtime/driver.py", line 20, in _initialize_obj
    self._obj = self._init_fn()
  File "/root/miniconda3/envs/1/lib/python3.10/site-packages/triton/runtime/driver.py", line 8, in _create_driver
    raise RuntimeError(f"{len(actives)} active drivers ({actives}). There should only be one.")
RuntimeError: 0 active drivers ([]). There should only be one.

[hiyouga]

pip uninstall galore_torch 试试

[shadow01a]

谢谢了，这确实有用！