add Readme

Author: metaphysicser

README.md

@@ -1,15 +1,6 @@
-# MLP-Communicator
+# PS-Mixer
 
-We begin our work on the basis of [MISA](https://github.com/declare-lab/MISA) initially, so the whole code architecture is similar to it, including the data process, data loader and evaluation metrics. Thanks to their open source spirit for saving us a lot of time.
-
-```
-@article{hazarika2020misa,
-  title={MISA: Modality-Invariant and-Specific Representations for Multimodal Sentiment Analysis},
-  author={Hazarika, Devamanyu and Zimmermann, Roger and Poria, Soujanya},
-  journal={arXiv preprint arXiv:2005.03545},
-  year={2020}
-}
-```
+![image-20220613001706944](https://jhfaoisehoiew.oss-cn-beijing.aliyuncs.com/img/image-20220613001706944.png)
 
 ## Requirements
 
@@ -36,6 +27,19 @@ pip install requirements.txt
 
 The repository is updating...
 
-### Contact
+## Acknowledgements
+
+We begin our work on the basis of [MISA](https://github.com/declare-lab/MISA) initially, so the whole code architecture is similar to it, including the data process, data loader and evaluation metrics. Thanks to their open source spirit for saving us a lot of time.
+
+```
+@article{hazarika2020misa,
+  title={MISA: Modality-Invariant and-Specific Representations for Multimodal Sentiment Analysis},
+  author={Hazarika, Devamanyu and Zimmermann, Roger and Poria, Soujanya},
+  journal={arXiv preprint arXiv:2005.03545},
+  year={2020}
+}
+```
+
+## Contact
 
 For any questions, please email at [zpl010720@gmail.com](zpl010720@gmail.com)

config.py

@@ -9,7 +9,7 @@
 word_emb_path = ''
 assert(word_emb_path is not None)
 
-project_dir = Path(__file__).resolve().parent
+project_dir = Path(__file__).resolve().parent.parent
 sdk_dir = project_dir.joinpath('CMU-MultimodalSDK')
 data_dir = project_dir.joinpath('datasets')
 data_dict = {'mosi': data_dir.joinpath('MOSI'), 'mosei': data_dir.joinpath(
@@ -100,7 +100,7 @@ def get_config(parse=True, **optional_kwargs):
     parser.add_argument('--test_duration', type=int, default=1)
 
     # Data
-    parser.add_argument('--data', type=str, default='mosi')  # mosi or mosei
+    parser.add_argument('--data', type=str, default='mosei')  # mosi or mosei
 
     # Parse arguments
     if parse:
@@ -116,7 +116,7 @@ def get_config(parse=True, **optional_kwargs):
     elif kwargs.data == "mosei":
         kwargs.num_classes = 1
         kwargs.batch_size = 64
-        kwargs.depth = 2
+        kwargs.depth = 1
     else:
         print("No dataset mentioned")
         exit()

train.py

@@ -11,7 +11,7 @@
 if __name__ == '__main__':
     # Setting random seed
     random_name = str(random())
-    random_seed = 5546
+    random_seed = 55654
     torch.manual_seed(random_seed)
     torch.cuda.manual_seed_all(random_seed)
     torch.backends.cudnn.deterministic = True