许多与 Glunts 的兼容性问题

Question

I would like to install the gluonts package. The problem here is that there are extremely many compatibility issues. I cannot manually solve each of these compatibility problems by trial and error. Is there a general way to solve the whole thing automatically?

I am working in the PyCharm environment with Pyton. to use a virtual environment I am using Anaconda.

The packages I need in any case are Jupyter, Numpy, Pandas, mxnet, matplotlib and just glounts.

I have also tried to install the whole thing in a different order, that didn't work either, so I installed gluonts first, which also worked, but when I then wanted to install mxnet is then the problem. Since I'm writing my bachelor thesis and accordingly under time pressure, I would be very grateful for your help.

My plan is to implement the whole thing in Pyton 3.6, I've also tried switching to Pyton 3.7, but that didn't help either. Now I will try it with Python 3.5.

When solving the problem, it should be noted that I am not completely free in my way of working, since I work on a unicomputer and do not have all the rights, etc. EDIT So I have now found a solution myself, how to use everything reasonably. You create with anaconda prompt a new environment and use there Python 3.9, when creating give the following code conda create -n env python=3.9 numpy=1.16.6 conda activate env pip install mxnet gluonts jupyter conda deactivate

now you could select this as normal environment in PyCharm. You could also look at the whole thing normally in the anaconda navigator and see which packages you have on it.

Answer 1

The general form is:

pip install gluonts

You should try this:

# support for mxnet models, faster datasets
pip install gluonts[mxnet,pro]  

# support for torch models, faster datasets
pip install gluonts[torch,pro]

and then try this example:

from gluonts.dataset.util import to_pandas
from gluonts.dataset.pandas import PandasDataset
from gluonts.dataset.repository.datasets import get_dataset
from gluonts.model.deepar import DeepAREstimator
from gluonts.mx import Trainer

dataset = get_dataset("airpassengers")

deepar = DeepAREstimator(prediction_length=12, freq="M", trainer=Trainer(epochs=5))
model = deepar.train(dataset.train)

# Make predictions
true_values = to_pandas(list(dataset.test)[0])
true_values.to_timestamp().plot(color="k")

prediction_input = PandasDataset([true_values[:-36], true_values[:-24], true_values[:-12]])
predictions = model.predict(prediction_input)

for color, prediction in zip(["green", "blue", "purple"], predictions):
    prediction.plot(color=f"tab:{color}")

plt.legend(["True values"], loc="upper left", fontsize="xx-large")

Here is the link to the gluonts 0.10.2 module:

https://pypi.org/project/gluonts/

Please state any errors that you see with this.