implement adding random noise to train demonstrations

gscan_metaseq2seq/util/dataset.py

@@ -1,7 +1,7 @@
 import numpy as np
 import torch
 from torch.utils.data import Dataset, IterableDataset
-
+import random
 from .padding import pad_to, recursive_pad_array
 
 
@@ -69,3 +69,34 @@ def __getitem__(self, i):
             self.indices = torch.randperm(len(self.dataset))
 
         return self.dataset[self.indices[i]]
+
+
+class AddRandomNoiseDataset(Dataset):
+    def __init__(self, dataset, ACTION2IDX, prob=0.01):
+        super().__init__()
+        self.dataset = dataset
+        self.ACTION2IDX = ACTION2IDX
+        self.prob = prob
+
+    def __len__(self):
+        return len(self.dataset)
+
+    def __getitem__(self, i):
+        item = self.dataset[i]
+
+        rand = random.uniform(0, 1)
+
+        def add_random_noise(instruction):
+            if len(np.where(instruction == self.ACTION2IDX["walk"])[0]) > 0:
+                walk_index = random.choice(np.where(instruction == self.ACTION2IDX["walk"])[0])
+                instruction[walk_index] = self.ACTION2IDX["turn left"]
+
+                return instruction
+
+            return instruction
+
+        if rand <= self.prob:
+            return item[0], add_random_noise(item[1]), item[2]
+
+        return item
+

gscan_metaseq2seq/util/logging.py

@@ -2,7 +2,7 @@
 import os
 
 from pytorch_lightning.loggers import CSVLogger
-from pytorch_lightning.loggers.logger import rank_zero_experiment
+from pytorch_lightning.loggers.base import rank_zero_experiment
 
 
 def get_most_recent_version(experiment_dir):

requirements.txt

@@ -1,13 +1,69 @@
-matplotlib>=3.3.3
-seaborn>=0.9.0
-h5py>=3.7.0
-pandas>=1.1.5
-positional-encodings>=4.0
-pytorch-lightning>=1.1.7
-rliable>=1.0.0
-scikit-learn>=0.24.1
-scipy>=1.7.0
-torch>=1.8.0
-torchvision>=0.11.2
-tqdm>=4.57.0
-torchmetrics>=0.4.0
+absl-py==1.4.0
+aiohttp==3.8.3
+aiosignal==1.3.1
+arch==5.3.0
+async-timeout==4.0.2
+asynctest==0.13.0
+attrs==22.2.0
+cachetools==5.3.0
+certifi==2022.12.7
+charset-normalizer==2.1.1
+cloudpickle==2.2.1
+cycler==0.11.0
+fonttools==4.38.0
+frozenlist==1.3.3
+fsspec==2023.1.0
+future==0.18.3
+google-auth==2.16.0
+google-auth-oauthlib==0.4.6
+grpcio==1.51.1
+gym==0.26.2
+gym-notices==0.0.8
+h5py==3.8.0
+idna==3.4
+importlib-metadata==6.0.0
+joblib==1.2.0
+kiwisolver==1.4.4
+Markdown==3.4.1
+MarkupSafe==2.1.2
+matplotlib==3.5.3
+multidict==6.0.4
+numpy==1.21.6
+oauthlib==3.2.2
+packaging==23.0
+pandas==1.3.5
+patsy==0.5.3
+Pillow==9.4.0
+positional-encodings==6.0.1
+property-cached==1.6.4
+protobuf==3.20.3
+pyasn1==0.4.8
+pyasn1-modules==0.2.8
+pyDeprecate==0.3.1
+pyparsing==3.0.9
+python-dateutil==2.8.2
+pytorch-lightning==1.5.10
+pytz==2022.7.1
+PyYAML==6.0
+requests==2.28.2
+requests-oauthlib==1.3.1
+rliable==1.0.8
+rsa==4.9
+scikit-learn==1.0.2
+scipy==1.7.3
+seaborn==0.12.2
+six==1.16.0
+statsmodels==0.13.5
+tensorboard==2.11.2
+tensorboard-data-server==0.6.1
+tensorboard-plugin-wit==1.8.1
+threadpoolctl==3.1.0
+torch==1.7.1
+torchmetrics==0.10.3
+torchvision==0.8.2
+tqdm==4.64.1
+typing_extensions==4.4.0
+urllib3==1.26.14
+Werkzeug==2.2.2
+yarl==1.8.2
+zipp==3.12.0

scripts/train_transformer.py

@@ -12,7 +12,7 @@
 from pytorch_lightning.loggers import TensorBoardLogger
 
 from gscan_metaseq2seq.models.embedding import BOWEmbedding
-from gscan_metaseq2seq.util.dataset import PaddingDataset, ReshuffleOnIndexZeroDataset
+from gscan_metaseq2seq.util.dataset import PaddingDataset, ReshuffleOnIndexZeroDataset, AddRandomNoiseDataset
 from gscan_metaseq2seq.util.load_data import load_data
 from gscan_metaseq2seq.util.logging import LoadableCSVLogger
 from gscan_metaseq2seq.util.scheduler import transformer_optimizer_config
@@ -51,7 +51,6 @@ def __init__(
                 nhead=nhead,
                 dim_feedforward=embedding_dim * 4,
                 dropout=dropout_p,
-                norm_first=norm_first,
             ),
             num_layers=nlayers,
         )
@@ -117,7 +116,6 @@ def __init__(
                 dim_feedforward=hidden_size * 4,
                 dropout=dropout_p,
                 nhead=nhead,
-                norm_first=norm_first,
             ),
             num_layers=nlayers,
         )
@@ -419,6 +417,8 @@ def main():
         args.train_demonstrations, args.valid_demonstrations_directory, args.dictionary
     )
 
+    train_demonstrations = AddRandomNoiseDataset(train_demonstrations, ACTION2IDX)
+
     IDX2WORD = {i: w for w, i in WORD2IDX.items()}
     IDX2ACTION = {i: w for w, i in ACTION2IDX.items()}