still getting the same error.

[koutkout]

I tried different lightning versions and it didn't work. I rolled back to 0.4.2 and although generation was running fine but training wasn't running!
I now have aitextgen 0.6.0 with pytorch-lightning 1.7.0
LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]
Traceback (most recent call last):
File "trainai[trainai.py](http://trainai.py/)", line 28, in
ai.train(data, batch_size=8, num_steps=50000, generate_every=5000)
File "C:\Python310\lib\site-packages\aitextgen[aitextgen.py](http://aitextgen.py/)", line 752, in train
trainer.fit(train_model)
File "C:\Python310\lib\site-packages\pytorch_lightning\trainer[trainer.py](http://trainer.py/)", line 700, in fit
self._call_and_handle_interrupt(
File "C:\Python310\lib\site-packages\pytorch_lightning\trainer[trainer.py](http://trainer.py/)", line 654, in _call_and_handle_interrupt
return trainer_fn(*args, **kwargs)
File "C:\Python310\lib\site-packages\pytorch_lightning\trainer[trainer.py](http://trainer.py/)", line 741, in _fit_impl
results = self._run(model, ckpt_path=self.ckpt_path)
File "C:\Python310\lib\site-packages\pytorch_lightning\trainer[trainer.py](http://trainer.py/)", line 1147, in _run
self.strategy.setup(self)
File "C:\Python310\lib\site-packages\pytorch_lightning\strategies[single_device.py](http://single_device.py/)", line 74, in setup
super().setup(trainer)
File "C:\Python310\lib\site-packages\pytorch_lightning\strategies[strategy.py](http://strategy.py/)", line 153, in setup
self.setup_optimizers(trainer)
File "C:\Python310\lib\site-packages\pytorch_lightning\strategies[strategy.py](http://strategy.py/)", line 141, in setup_optimizers
self.optimizers, self.lr_scheduler_configs, self.optimizer_frequencies = _init_optimizers_and_lr_schedulers(
File "C:\Python310\lib\site-packages\pytorch_lightning\core[optimizer.py](http://optimizer.py/)", line 194, in _init_optimizers_and_lr_schedulers
_validate_scheduler_api(lr_scheduler_configs, model)
File "C:\Python310\lib\site-packages\pytorch_lightning\core[optimizer.py](http://optimizer.py/)", line 351, in _validate_scheduler_api
raise MisconfigurationException(
pytorch_lightning.utilities.exceptions.MisconfigurationException: The provided lr scheduler LambdaLR doesn't follow PyTorch's LRScheduler API. You should override the LightningModule.lr_scheduler_step hook with your own logic if you are using a custom LR scheduler.

=============
if i upgrade to lightning==2.0.0 then I will get:
Traceback (most recent call last):
File "trainai[trainai.py](http://trainai.py/)", line 1, in
from aitextgen.TokenDataset import TokenDataset
File "C:\Python310\lib\site-packages\aitextgen[init.py](http://init.py/)", line 1, in
from .aitextgen import aitextgen # noqa
File "C:\Python310\lib\site-packages\aitextgen[aitextgen.py](http://aitextgen.py/)", line 14, in
from pytorch_lightning.plugins import DeepSpeedPlugin
ImportError: cannot import name 'DeepSpeedPlugin' from 'pytorch_lightning.plugins' (C:\Python310\lib\site-packages\pytorch_lightning\plugins_init_.py)

Please help.

[astrolul]

Same I recently tried this out and I was getting all sorts of errors, let me know when this is fixed because I want to use it ASAP.

[analyticray]

I manage to resolve the error on the google colab by running this:

!pip install -qq pytorch-lightning==1.7.0 transformers==4.21.3 aitextgen==0.6.0

Please do let me know if it also solves your issue on colab

[koutkout]

Thanks. It was solved earlier. But now there's another problem. I remember testing it a week ago and it was working but for unknown reason it's not working now.
ImportError Traceback (most recent call last)
in <cell line: 1>()
----> 1 from aitextgen import aitextgen
2 from aitextgen.colab import mount_gdrive, copy_file_from_gdrive
3 ai = aitextgen(tf_gpt2="124M", to_gpu=True)

1 frames
/usr/local/lib/python3.9/dist-packages/aitextgen/aitextgen.py in
12 import torch
13 from pkg_resources import resource_filename
---> 14 from pytorch_lightning.plugins import DeepSpeedPlugin
15 from tqdm.auto import trange
16 from transformers import (

ImportError: cannot import name 'DeepSpeedPlugin' from 'pytorch_lightning.plugins' (/usr/local/lib/python3.9/dist-packages/pytorch_lightning/plugins/init.py)

[scorixear]

Thanks. It was solved earlier. But now there's another problem. I remember testing it a week ago and it was working but for unknown reason it's not working now. ImportError Traceback (most recent call last) in <cell line: 1>() ----> 1 from aitextgen import aitextgen 2 from aitextgen.colab import mount_gdrive, copy_file_from_gdrive 3 ai = aitextgen(tf_gpt2="124M", to_gpu=True)

1 frames /usr/local/lib/python3.9/dist-packages/aitextgen/aitextgen.py in 12 import torch 13 from pkg_resources import resource_filename ---> 14 from pytorch_lightning.plugins import DeepSpeedPlugin 15 from tqdm.auto import trange 16 from transformers import (

ImportError: cannot import name 'DeepSpeedPlugin' from 'pytorch_lightning.plugins' (/usr/local/lib/python3.9/dist-packages/pytorch_lightning/plugins/init.py)

as you already figured out, aitextgen is build on older versions of pytorch_lightning, but doesn't have any max-version in their requirements.
#216 updates those library changes, they are however mostly for show, the algorithms behind them (or api calls) all stayed the same.