Neuralforecasting pytorch dependency issues

[egolub22]

I am attempting to utilize the neuralforecasting package as part of my ml solution but I am getting a series of pytorch related dependency errors that are preventing me from using it. The first of these errors appears when I try to initially import the package.

`
Import statements:

from neuralforecast import NeuralForecast
from neuralforecast.models import LSTM, NHITS, RNN
from neuralforecast.auto import AutoNHITS
from neuralforecast.losses.pytorch import MAE
from neuralforecast.losses.pytorch import MAPE

import plotly
from plotly.subplots import make_subplots
import plotly.graph_objs as go
from utilsforecast.losses import *

print(version('plotly'))
print(torch.__version__)

`

Error:

I fixed this using the terminal comand: pip install neuralforecast "torch<2.0.0", which was suggested in another online forum. I am not sure if this is the source of the following issues. The next issue I encounter occures when I try to instantiate an mlforecast object.

`
Code:

horizon = 31
models = [LSTM(h=horizon,                    # Forecast horizon
               max_steps=500,                # Number of steps to train
               scaler_type='standard',       # Type of scaler to normalize data
               encoder_hidden_size=64,       # Defines the size of the hidden state of the LSTM
               decoder_hidden_size=64,),     # Defines the number of hidden units of each layer of the MLP decoder
          NHITS(h=horizon,                   # Forecast horizon
                input_size=2 * horizon,      # Length of input sequence
                max_steps=100,               # Number of steps to train
                n_freq_downsample=[2, 1, 1]) # Downsampling factors for each stack output
          ]
nf = NeuralForecast(models=models, freq='D')
nf.fit(df=df)

`

Error:

I'm not sure if there is a certain version of torch/pytorch that is recommended for the neuralforecast package but I have tried upgrading those packages to various version as well as the neuralforcasting package itself to eliminate the error to no avail.

[jmoralez]

You need pytorch_lightning>=2.1.0

[egolub22]

upgrading pytorch_lightning to 2.1.0 still gives me the first error about disable_possible_user_warnings

[jmoralez]

It's there https://github.com/Lightning-AI/pytorch-lightning/blob/2.1.0/src/lightning/fabric/utilities/warnings.py#L26