From b7206d2d4c2a861d427f590919e3750ff18d55c8 Mon Sep 17 00:00:00 2001
From: Seyong Kim <seyong.dev@gmail.com>
Date: Sun, 5 May 2024 15:54:26 +0900
Subject: [PATCH] Update README.md

---
 README.md | 60 ++++++++++++++++++++++++++++---------------------------
 1 file changed, 31 insertions(+), 29 deletions(-)

diff --git a/README.md b/README.md
index c393c88..7db03f7 100644
--- a/README.md
+++ b/README.md
@@ -4,8 +4,12 @@
 This repository contains the implementation of a bidirectional Convolutional LSTM (ConvLSTM) in PyTorch, as described in the paper [Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting](https://arxiv.org/abs/1506.04214). The ConvLSTM model is particularly useful for spatiotemporal predictions where both spatial and temporal dynamics need to be captured.
 
 ## Features
-- Implementation of ConvLSTM with bidirectional capability.
-- Suitable for tasks like video frame prediction, anomaly detection in videos, and other spatiotemporal sequence prediction tasks.
+- Implements bidirectional ConvLSTM, allowing the model to capture both forward and backward temporal dependencies
+- Supports batch processing and variable sequence lengths
+- Provides configurable parameters for input channels, hidden channels, kernel size, padding, stride, and bias
+- Includes a custom `ConvGate` module for efficient computation of gate activations
+- Compatible with PyTorch's `nn.Module` and can be easily integrated into existing PyTorch models
+
 
 ## Environment Setup
 - Python 3.8
@@ -13,7 +17,6 @@ This repository contains the implementation of a bidirectional Convolutional LST
 
 To set up your environment to run the code, you can follow these steps:
 ```
-bash
 conda create --name convlstm python=3.8
 conda activate convlstm
 pip install torch==1.13 torchvision
@@ -21,32 +24,31 @@ pip install torch==1.13 torchvision
 
 ### Basic Example
 ```
-from model import ConvLSTM
-import torch
-
-# Example parameters
-input_channels = 3
-hidden_channels = 16
-kernel_size = (3, 3)
-padding = (1, 1)
-stride = (1, 1)
-bias = True
-batch_first = True
-bidirectional = True
-
-model = ConvLSTM(
-    in_channels=input_channels,
-    hidden_channels=hidden_channels,
-    kernel_size=kernel_size,
-    padding=padding,
-    stride=stride,
-    bias=bias,
-    batch_first=batch_first,
-    bidirectional=bidirectional
+from bidirectional_conv_lstm import ConvLSTM
+
+# Create a ConvLSTM instance
+conv_lstm = ConvLSTM(
+    in_channels=1,
+    hidden_channels=64,
+    kernel_size=3,
+    padding=1,
+    stride=1,
+    bias=True,
+    batch_first=True,
+    bidirectional=True
 )
 
-# Example input tensor
-batch_size, seq_len, height, width = 1, 10, 64, 64
-x = torch.rand(batch_size, seq_len, input_channels, height, width)
-output, (hidden_state, cell_state) = model(x)
+# Forward pass
+input_tensor = torch.randn(batch_size, seq_len, in_channels, height, width)
+output, (hidden_state, cell_state) = conv_lstm(input_tensor)
 ```
+
+### References
+
+Shi, X., Chen, Z., Wang, H., Yeung, D. Y., Wong, W. K., & Woo, W. C. (2015). Convolutional LSTM network: A machine learning approach for precipitation nowcasting. In Advances in neural information processing systems (pp. 802-810). [arXiv:1506.04214]
+
+### Contributing
+Contributions to this repository are welcome! If you have any suggestions, bug reports, or feature requests, please open an issue or submit a pull request.
+
+### License
+This project is licensed under the MIT License.