ourownstory / ar-net Goto Github PK
View Code? Open in Web Editor NEWA simple Auto-Regressive Neural Network for time-series
License: MIT License
A simple Auto-Regressive Neural Network for time-series
License: MIT License
more like a questions about the feature instead of issue, I am trying to understand if it is possible to use multivariate TS dataset for forecasting using AR-Net. In case yes, are there some examples?
Also, can it take care of dependencies b.w. variables?
I was trying to use the AR-Net for one of my project when I came up with the following error:
`ValueError Traceback (most recent call last)
in
----> 1 m.fit(deltas)
~/opt/anaconda3/lib/python3.8/site-packages/arnet/ar_net.py in fit(self, series, plot)
217 self.make_datasets(series)
218 self.create_learner()
--> 219 self.fit_one_cycle(plot=plot)
220 return self
221
~/opt/anaconda3/lib/python3.8/site-packages/arnet/ar_net.py in fit_one_cycle(self, n_epoch, lr, cycles, plot)
201
202 if lr is None:
--> 203 self.find_lr(plot=plot)
204 lr = self.lr
205 for i in range(0, cycles):
~/opt/anaconda3/lib/python3.8/site-packages/arnet/ar_net.py in find_lr(self, plot)
187 if self.learn is None:
188 raise ValueError("create learner first.")
--> 189 lr_at_min, lr_steep = self.learn.lr_find(start_lr=1e-6, end_lr=10, num_it=300, show_plot=plot)
190 if plot:
191 plt.show()
ValueError: not enough values to unpack (expected 2, got 1)`
Apparently the Fast-AI functionality was changed: https://forums.fast.ai/t/error-in-lr-find/89968
So the code requires a small update so that the AR-Net can work.
Hey, don't know wether this is actually an issue with the module or just user error, but I'm getting a Name Error and I honestly can't figure out why.
Any idea what could cause this?
Cell In[1], [line 1](vscode-notebook-cell:?execution_count=1&line=1)
----> [1](vscode-notebook-cell:?execution_count=1&line=1) import arnet
File .\arnet\__init__.py:24
[21](./arnet/__init__.py:21) from fastai.basics import *
[22](./arnet/__init__.py:22) from fastai.tabular.all import *
---> [24](./arnet/__init__.py:24) from .ar_net import ARNet
[26](.arnet/__init__.py:26) # from .ar_net_legacy import init_ar_learner
[27](./arnet/__init__.py:27) from .utils_data import load_from_file, tabularize_univariate, estimate_noise, split_by_p_valid
File .\arnet\ar_net.py:13
[11](./arnet/ar_net.py:11) from fastai.data.transforms import Normalize
[12](./arnet/ar_net.py:12) from fastai.learner import load_learner
---> [13](./arnet/ar_net.py:13) from fastai.distributed import ParallelTrainer
[15](./arnet/ar_net.py:15) from arnet import utils, utils_data, plotting
[16](./arnet/ar_net.py:16) from arnet.fastai_mods import SparsifyAR, huber, sTPE, get_loss_func
File .\fastai\distributed.py:139
[136](./fastai/distributed.py:136) _hidden_params = ["mixed_precision", "fp16", "log_with", "logging_dir", "step_scheduler_with_optimizer"]
[138](./fastai/distributed.py:138) # %% ../nbs/20a_distributed.ipynb 30
--> [139](./fastai/distributed.py:139) class DistributedTrainer(Callback):
[140](./fastai/distributed.py:140) "Wrap `model` in `DistributedDataParallel` and `dls` in `DistributedDL`"
[141](./fastai/distributed.py:141) order = 11
...
[146](./fastai/distributed.py:146) ):
[147](./fastai/distributed.py:147) store_attr()
[148](./fastai/distributed.py:148) self.accelerator = Accelerator(**kwargs)
NameError: name 'Accelerator' is not defined
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.