update gluonts version

[kashif]

update to use latest gluonts version

[ashok-arjun]

This also requires latest pandas versions right?

[ashok-arjun]

It'll be nice to check if the Colab works (with the current requirements file) before we merge this

[kashif]

@kiranmaya I've pinned the versions of numpy and pandas as was suggested... can you test if that works?

[kiranmaya]

It was working now. will double-check and let you know.

[kashif]

thanks... i'll wait for your confirmation and merge

[kiranmaya]

working when numpy is at v 1.26.4 . but when I run demo note book as it is ,it wont execute ,because somehow ,numpy is at 2.1.0 .
maybe collab comes with default packages by default numpy is at 2.1.0.
I print out the working versions with !pip list
packages.txt

[kashif]

@kiranmaya can we pin numpy to <2 ?

[kiranmaya]

see my packages list once .
packages.txt

[kashif]

ok lets pin numpy to <= 1.26.4

[kiranmaya]

ok lets pin numpy to <= 1.26.4
with
gluonts[torch]
numpy<=1.26.4
torch>=2.0.0
wandb
scipy
pandas==2.2.3
huggingface_hub[cli]
matplotlib
i got below erro in fresh notebook run .
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
google-colab 1.0.0 requires pandas==2.2.2, but you have pandas 2.2.3 which is incompatible.
torchaudio 2.6.0+cu124 requires torch==2.6.0, but you have torch 2.7.0 which is incompatible.
torchvision 0.21.0+cu124 requires torch==2.6.0, but you have torch 2.7.0 which is incompatible.
fastai 2.7.19 requires torch<2.7,>=1.10, but you have torch 2.7.0 which is incompatible.
thinc 8.3.6 requires numpy<3.0.0,>=2.0.0, but you have numpy 1.26.4 which is incompatible.

ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

[kashif]

or lets not change numpy and see ?

[kiranmaya]

or lets not change numpy and see ?

i think ,torch needs to be updated to 2.5.1 , currently it is torch>=2.0.0 so pip went all the way to top .
i suggest creating a new environment for this project, since lag-llama is not a package, this needs to separate environment-based project.

[kashif]

@kiranmaya i just tested and all seems to work with pytorch 2.6.0 and 2.7.0

[kiranmaya]

just one dependency is enough,gluonts[torch] , it is installing whatever it needs.
!pip install gluonts[torch]

[kashif]

ok let me remove the rest