Describe the bug
Looks like the package versions in requirements.txt files are too
loose.
First, got error about NeptuneLogger not being found. And further
errors even if that is ignored.
Steps/Code to reproduce bug
pip install torch torchvision torchaudio
pip install packaging
pip install huggingface_hub==0.23.2
git clone https://github.com/AI4Bharat/NeMo.git
cd NeMo/
bash reinstall.sh
python3 hello.py
where hello.py is like:
import nemo.collections.asr as nemo_asr
import torch
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model_path = "/home/username/indicconformer_stt_ml_hybrid_rnnt_large.nemo"
model = nemo_asr.models.EncDecCTCModelBPE.restore_from(restore_path=model_path)
model.freeze()
model = model.to(device)
Error:
$ python3 hello.py
[NeMo W 2026-08-06 11:44:59 nemo_logging:349] /home/username/venvs/nemo-py310/lib/python3.10/site-packages/megatron/core/tensor_parallel/layers.py:220: FutureWarning: `torch.cuda.amp.custom_fwd(args...)` is deprecated. Please use `torch.amp.custom_fwd(args..., device_type='cuda')` instead.
def forward(ctx, input, weight, bias, gradient_accumulation_fusion,
[NeMo W 2026-08-06 11:44:59 nemo_logging:349] /home/username/venvs/nemo-py310/lib/python3.10/site-packages/megatron/core/tensor_parallel/layers.py:250: FutureWarning: `torch.cuda.amp.custom_bwd(args...)` is deprecated. Please use `torch.amp.custom_bwd(args..., device_type='cuda')` instead.
def backward(ctx, grad_output):
Traceback (most recent call last):
File "/media/username/NeMo/hello.py", line 2, in <module>
import nemo.collections.asr as nemo_asr
File "/media/username/NeMo/nemo/collections/asr/__init__.py", line 15, in <module>
from nemo.collections.asr import data, losses, models, modules
File "/media/username/NeMo/nemo/collections/asr/models/__init__.py", line 15, in <module>
from nemo.collections.asr.models.aed_multitask_models import EncDecMultiTaskModel
File "/media/username/NeMo/nemo/collections/asr/models/aed_multitask_models.py", line 29, in <module>
from nemo.collections.asr.metrics import BLEU, WER
File "/media/username/NeMo/nemo/collections/asr/metrics/__init__.py", line 15, in <module>
from nemo.collections.asr.metrics.bleu import BLEU
File "/media/username/NeMo/nemo/collections/asr/metrics/bleu.py", line 22, in <module>
from nemo.collections.asr.parts.submodules.multitask_decoding import AbstractMultiTaskDecoding
File "/media/username/NeMo/nemo/collections/asr/parts/submodules/multitask_decoding.py", line 23, in <module>
from nemo.collections.asr.parts.submodules.multitask_beam_decoding import (
File "/media/username/NeMo/nemo/collections/asr/parts/submodules/multitask_beam_decoding.py", line 21, in <module>
from nemo.collections.asr.modules.transformer import BeamSearchSequenceGenerator
File "/media/username/NeMo/nemo/collections/asr/modules/__init__.py", line 15, in <module>
from nemo.collections.asr.modules.audio_modules import (
File "/media/username/NeMo/nemo/collections/asr/modules/audio_modules.py", line 21, in <module>
from nemo.collections.asr.modules.conformer_encoder import ConformerEncoder
File "/media/username/NeMo/nemo/collections/asr/modules/conformer_encoder.py", line 26, in <module>
from nemo.collections.asr.models.configs import CacheAwareStreamingConfig
File "/media/username/NeMo/nemo/collections/asr/models/configs/__init__.py", line 15, in <module>
from nemo.collections.asr.models.configs.asr_models_config import (
File "/media/username/NeMo/nemo/collections/asr/models/configs/asr_models_config.py", line 27, in <module>
from nemo.core.config import modelPT as model_cfg
File "/media/username/NeMo/nemo/core/config/modelPT.py", line 22, in <module>
from nemo.utils import exp_manager
File "/media/username/NeMo/nemo/utils/exp_manager.py", line 35, in <module>
from pytorch_lightning.loggers import MLFlowLogger, NeptuneLogger, TensorBoardLogger, WandbLogger
ImportError: cannot import name 'NeptuneLogger' from 'pytorch_lightning.loggers' (/home/username/venvs/nemo-py310/lib/python3.10/site-packages/pytorch_lightning/loggers/__init__.py)
Expected behavior
No error.
Environment overview (please complete the following information)
- Local installation via venv and pip
Commands used for installation mentioned above.
Environment details
If NVIDIA docker image is used you don't need to specify these.
Otherwise, please provide:
- OS: Debian 13
- PyTorch version: as provided in requirements.txt
- Python version: 3.10.16
pytorch-lightning==2.6.5
torch==2.13.0
torchaudio==2.11.0
torchdiffeq==0.2.5
torchmetrics==1.9.0
torchsde==0.2.6
torchvision==0.28.0
Additional context
Integrated GPU. Not NVIDIA.
Describe the bug
Looks like the package versions in requirements.txt files are too
loose.
First, got error about NeptuneLogger not being found. And further
errors even if that is ignored.
Steps/Code to reproduce bug
where hello.py is like:
Error:
Expected behavior
No error.
Environment overview (please complete the following information)
Commands used for installation mentioned above.
Environment details
If NVIDIA docker image is used you don't need to specify these.
Otherwise, please provide:
Additional context
Integrated GPU. Not NVIDIA.