64250d0938
Running examples/cli/gmlvq.py crashes with:
```
---------------------------------------------------------------------------
KeyError Traceback (most recent call last)
~/miniconda3/envs/py38/lib/python3.8/site-packages/pytorch_lightning/utilities/parsing.py in __getattr__(self, key)
249 try:
--> 250 return self[key]
251 except KeyError as exp:
KeyError: 'distribution'
The above exception was the direct cause of the following exception:
AttributeError Traceback (most recent call last)
~/work/repos/prototorch_models/examples/cli/gmlvq.py in <module>
10
11
---> 12 cli = LightningCLI(GMLVQMNIST)
~/miniconda3/envs/py38/lib/python3.8/site-packages/pytorch_lightning/utilities/cli.py in __init__(self, model_class, datamodule_class, save_config_callback, trainer_class, trainer_defaults, seed_everything_default, description, env_prefix, env_parse, parser_kwargs, subclass_mode_model, subclass_mode_data)
168 seed_everything(self.config['seed_everything'])
169 self.before_instantiate_classes()
--> 170 self.instantiate_classes()
171 self.prepare_fit_kwargs()
172 self.before_fit()
~/miniconda3/envs/py38/lib/python3.8/site-packages/pytorch_lightning/utilities/cli.py in instantiate_classes(self)
211 self.config_init = self.parser.instantiate_subclasses(self.config)
212 self.instantiate_datamodule()
--> 213 self.instantiate_model()
214 self.instantiate_trainer()
215
~/miniconda3/envs/py38/lib/python3.8/site-packages/pytorch_lightning/utilities/cli.py in instantiate_model(self)
228 self.model = self.config_init['model']
229 else:
--> 230 self.model = self.model_class(**self.config_init.get('model', {}))
231
232 def instantiate_trainer(self) -> None:
~/work/repos/prototorch_models/prototorch/models/glvq.py in __init__(self, hparams, **kwargs)
307 def __init__(self, hparams, **kwargs):
308 distance_fn = kwargs.pop("distance_fn", omega_distance)
--> 309 super().__init__(hparams, distance_fn=distance_fn, **kwargs)
310 omega = torch.randn(self.hparams.input_dim,
311 self.hparams.latent_dim,
~/work/repos/prototorch_models/prototorch/models/glvq.py in __init__(self, hparams, **kwargs)
39 # Layers
40 self.proto_layer = LabeledComponents(
---> 41 distribution=self.hparams.distribution,
42 initializer=self.prototype_initializer(**kwargs))
43
~/miniconda3/envs/py38/lib/python3.8/site-packages/pytorch_lightning/utilities/parsing.py in __getattr__(self, key)
250 return self[key]
251 except KeyError as exp:
--> 252 raise AttributeError(f'Missing attribute "{key}"') from exp
253
254 def __setattr__(self, key, val):
AttributeError: Missing attribute "distribution"
```
ProtoTorch Models
Pre-packaged prototype-based machine learning models using ProtoTorch and PyTorch-Lightning.
Installation
To install this plugin, simply run the following command:
pip install prototorch_models
The plugin should then be available for use in your Python environment as
prototorch.models.
Note: Installing the models plugin should automatically install a suitable version of ProtoTorch.
Available models
LVQ Family
- Learning Vector Quantization 1 (LVQ1)
- Generalized Learning Vector Quantization (GLVQ)
- Generalized Relevance Learning Vector Quantization (GRLVQ)
- Generalized Matrix Learning Vector Quantization (GMLVQ)
- Localized and Generalized Matrix Learning Vector Quantization (LGMLVQ)
- Limited-Rank Matrix Learning Vector Quantization (LiRaMLVQ)
- Learning Vector Quantization Multi-Layer Network (LVQMLN)
- Siamese GLVQ
- Cross-Entropy Learning Vector Quantization (CELVQ)
- Robust Soft Learning Vector Quantization (RSLVQ)
Other
- k-Nearest Neighbors (KNN)
- Neural Gas (NG)
- Growing Neural Gas (GNG)
Work in Progress
- Classification-By-Components Network (CBC)
- Learning Vector Quantization 2.1 (LVQ2.1)
Planned models
- Median-LVQ
- Generalized Tangent Learning Vector Quantization (GTLVQ)
- Probabilistic Learning Vector Quantization (PLVQ)
- Self-Incremental Learning Vector Quantization (SILVQ)
Development setup
It is recommended that you use a virtual environment for development. If you do
not use conda, the easiest way to work with virtual environments is by using
virtualenvwrapper. Once
you've installed it with pip install virtualenvwrapper, you can do the
following:
export WORKON_HOME=~/pyenvs
mkdir -p $WORKON_HOME
source /usr/local/bin/virtualenvwrapper.sh # location may vary
mkvirtualenv pt
Once you have a virtual environment setup, you can start install the models
plugin with:
workon pt
git clone git@github.com:si-cim/prototorch_models.git
cd prototorch_models
git checkout dev
pip install -e .[all] # \[all\] if you are using zsh or MacOS
Note: Please avoid installing Tensorflow in this environment.
To assist in the development process, you may also find it useful to install
yapf, isort and autoflake. You can install them easily with pip.
FAQ
How do I update the plugin?
If you have already cloned and installed prototorch and the
prototorch_models plugin with the -e flag via pip, all you have to do is
navigate to those folders from your terminal and do git pull to update.