Merge pull request #1 from moienr/dev-SSCL-SoilNet

Dev sscl soil net

Author: nafisehkakhani

.gitignore

@@ -1 +1,6 @@
-**/__pycache__/ 
+**/__pycache__/ 
+results_physics/*
+results_simclr/*
+results/*
+results_simclr_lucas/*
+*.tif

Accuracy_assessment.ipynb

@@ -83,7 +83,8 @@
    "outputs": [],
    "source": [
     "# Read the CSV file which is first setup: Base\n",
-    "df = pd.read_csv('...')\n",
+    "# df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\RUN_D_2023_06_02_T_20_57_Nafiseh.csv')\n",
+    "df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\onlytest_D_2023_06_06_T_09_56_Nafiseh.csv')\n",
     "\n",
     "y_true = df['y_real']\n",
     "y_pred = df['y_pred']\n",
@@ -106,8 +107,8 @@
    "outputs": [],
    "source": [
     "# Read the CSV file which is second setup: added spatial attention\n",
-    "# df = pd.read_csv('...')\n",
-    "df = pd.read_csv('...')\n",
+    "# df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\RUN_D_2023_06_03_T_09_41_Nafiseh.csv')\n",
+    "df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\onlytest_D_2023_06_06_T_09_44_Nafiseh.csv')\n",
     "\n",
     "y_true = df['y_real']\n",
     "y_pred = df['y_pred']\n",
@@ -130,8 +131,8 @@
    "outputs": [],
    "source": [
     "# Read the CSV file which is 3th setup: LSTM\n",
-    "# df = pd.read_csv('...')\n",
-    "df = pd.read_csv('...')\n",
+    "# df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\RUN_D_2023_06_05_T_16_02_Nafiseh.csv')\n",
+    "df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\onlytest_D_2023_06_05_T_16_02_Nafiseh.csv')\n",
     "\n",
     "y_true = df['y_real']\n",
     "y_pred = df['y_pred']\n",
@@ -172,8 +173,8 @@
    ],
    "source": [
     "# Read the CSV file which is RF\n",
-    "# df = pd.read_csv('...')\n",
-    "df = pd.read_csv('...')\n",
+    "# df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\RUN_D_2023_06_05_T_16_02_Nafiseh.csv')\n",
+    "df = pd.read_csv('C:\\\\Users\\\\nkakhani\\\\_Multimodal\\\\SoilNet-5\\\\SoilNet\\\\results\\\\RF_D_2023_06_06_T_13_27_Nafiseh.csv')\n",
     "\n",
     "y_true = df['y_true']\n",
     "y_pred = df['y_pred']\n",

MultimodalDL.docx

@@ -1,13 +1,5 @@
 # SoilNet
-A Spatio-temporal Framework for Soil Property Prediction with Digital Soil Mapping (DSM)
 
-This new architecture, incorporates spatial information using a base convolutional neural network (CNN) model and spatial attention mechanism, along with climate temporal information using a long short-term memory (LSTM) network. 
-
-In case you need more information, feel free to send an email: [email protected]
-
-Paper: https://arxiv.org/abs/2308.03586
-
-To access LUCAS topsoil dataset: https://esdac.jrc.ec.europa.eu/content/topsoil-physical-properties-europe-based-lucas-topsoil-data
 
 <!-- ## Experiments
 
@@ -26,185 +18,29 @@ To access LUCAS topsoil dataset: https://esdac.jrc.ec.europa.eu/content/topsoil-
 
 
 
+
+
+
+
 <!-- ### MODEL
 - ~~Add Spatial Attention Module~~
 - ~~FCNN + Regressor~~
 - ~~FCNN + SA + Regressor~~
 - ~~LSTM~~
  -->
- 
 <!-- ### DATASET
 - ~~Add TerraClimate Dataset~~
 - ~~Update the ClimateInformation.js (in processing)~~
-- ~~Write SRTM + Slope dataset donwlaoder~~ 
- ### Analysis
+- ~~Write SRTM + Slope dataset donwlaoder~~ -->
+
+### Analysis
 - CNN
 - CNN + Att 
 - CNN + Att + LSTM
 - RF with no timeseries data
-- RF + timeseries data -->
+- RF + timeseries data
 
 ---
-<!-- ![oc_all](https://github.com/moienr/SoilNet/blob/d0255c1ce411e631265daf311f1ca0d68b7b0412/readme_imgs/overallarc2.png) -->
----
- ![SoilNET](./readme_imgs/overallarc2.png)
+<!-- ![oc_germany](https://github.com/moienr/SoilNet/blob/da789f2bf2f2df5dacca6b44ea2be63ed516e54f/dataset/oc_histogram.png) -->
 ---
-
-<!-- ### 
-# Usage
-
-
-## Installation
-
-1. **Prerequisites**:
-   - Install Conda: Make sure you have Conda installed on your system.
-
-2. **Clone the Repository**:
-   ```bash
-   git clone https://github.com/moienr/SoilNet.git
-   ```
-
-3. **Create the Environments**:
-   ```bash
-    cd SoilNet
-    ```
-    3.1. for training:
-    ```bash  
-    conda env create -f requirements/pytorch_reqs.yml
-    ```
-
-    3.2. for dataset:
-    ```bash  
-    conda env create -f requirements/geemap_reqs.yml
-    ```
-
-4. **Activate the Environment**:
-
-    4.1. for train:
-
-    ```bash
-    conda activate pytorch
-    ```
-
-    4.2. to download the dataset:
-
-    ```bash
-    conda activate geemap
-    ```
-
-## Training
-
-5. **Run the Code**:
-
-    Flags are explained in the next section.
-
-    ```bash
-    python train.py -ne 100 -tbs 8 -ne 10 -ca resnet101
-    ```
-    
-    Although to train, you're gonna need to have the `.csv` files. namely LUCAS dataset under the flag of `--lucas_csv` and the TerraClimate dataset under the flag of `--climate_csv_folder_path`.
-
-The output is a **Training Plot and** a **JSON** file containing all of the results of the **cross-validation**. all will be saved in the `results/` folder
-
--->
-
-
-
-
-<!--
-## Explanation of Command-line Flags for `train.py`
-
-The `train.py` script accepts several command-line flags (arguments) that allow you to customize the training process. These flags help adjust various settings and parameters for the model training. Below is a detailed explanation of each flag:
-
-1. `-nw` or `--num_workers`:
-   - Type: Integer
-   - Default: 2
-   - Description: Number of workers used for data loading during training. Adjust this value based on your system's capabilities to optimize data loading efficiency.
-
-2. `-tbs` or `--train_batch_size`:
-   - Type: Integer
-   - Default: 4
-   - Description: Batch size used during training. A larger batch size can increase training speed but may require more memory.
-
-3. `-Tbs` or `--test_batch_size`:
-   - Type: Integer
-   - Default: 4
-   - Description: Batch size used during testing. Similar to the training batch size, it affects memory consumption during testing.
-
-4. `-lr` or `--learning_rate`:
-   - Type: Float
-   - Default: 0.0001
-   - Description: Learning rate used for the optimization algorithm during training.
-
-5. `-ne` or `--num_epochs`:
-   - Type: Integer
-   - Default: 2
-   - Description: Number of training epochs. An epoch is a complete pass through the entire training dataset.
-
-6. `-lrs` or `--lr_scheduler`:
-   - Choices: 'step', 'plateau', or None (case-sensitive)
-   - Default: 'step'
-   - Description: Learning rate scheduler type. Choose from 'step' (step-wise decay), 'plateau' (decay on validation loss plateau), or None (no learning rate decay).
-
-7. `-oc` or `--oc_max`:
-   - Type: Integer
-   - Default: 87
-   - Description: Maximum value for OC during training. Adjust this value based on your specific use case.
-
-8. `-us` or `--use_srtm`:
-   - Action: Store True
-   - Default: True
-   - Description: Enable or disable the use of SRTM (Shuttle Radar Topography Mission) data during training.
-
-9. `-usa` or `--use_spatial_attention`:
-   - Action: Store True
-   - Default: True
-   - Description: Enable or disable the use of spatial attention in the model architecture.
-
-10. `-ca` or `--cnn_architecture`:
-    - Choices: 'vgg16', 'resnet101' (case-sensitive)
-    - Default: 'vgg16'
-    - Description: Choose the CNN architecture for the model. Options are 'vgg16' or 'resnet101'.
-
-11. `-rv` or `--reg_version`:
-    - Type: Integer
-    - Default: 2
-    - Description: Regression version used during training. Adjust this value based on your specific use case.
-
-12. `-ulb` or `--use_lstm_branch`:
-    - Action: Store True
-    - Default: True
-    - Description: Enable or disable the use of the LSTM branch in the model architecture.
-
-13. `-tl8` or `--train_l8_folder_path`:
-    - Type: String
-    - Default: 'D:\python\SoilNet\dataset\l8_images\\train\\'
-    - Description: Path to the training L8 (Landsat 8) image folder.
-
-14. `-tsl8` or `--test_l8_folder_path`:
-    - Type: String
-    - Default: 'D:\python\SoilNet\dataset\l8_images\\test\\'
-    - Description: Path to the test L8 (Landsat 8) image folder.
-
-15. `-vl8` or `--val_l8_folder_path`:
-    - Type: String
-    - Default: 'D:\python\SoilNet\dataset\l8_images\\val\\'
-    - Description: Path to the validation L8 (Landsat 8) image folder.
-
-16. `-tvsl8` or `--testval_l8_folder_path`:
-    - Type: String
-    - Default: 'D:\python\SoilNet\dataset\l8_images\\val\\'
-    - Description: Path to the test/validation L8 (Landsat 8) image folder.
-
-17. `-lcp` or `--lucas_csv_path`:
-    - Type: String
-    - Default: 'D:\python\SoilNet\dataset\LUCAS_2015_all.csv'
-    - Description: Path to the LUCAS CSV file.
-
-18. `-ccp` or `--climate_csv_folder_path`:
-    - Type: String
-    - Default: "D:\\python\\SoilNet\\dataset\\Climate\\All\\filled\\"
-    - Description: Path to the climate CSV folder.
-
-These command-line flags allow you to configure various aspects of the model training process based on your specific requirements and dataset. Adjust the values according to your needs when running the `train.py` script.
--->
+![oc_all](https://github.com/moienr/SoilNet/blob/d0255c1ce411e631265daf311f1ca0d68b7b0412/readme_imgs/oc_all.png)

README.md

@@ -0,0 +1 @@
+**/*.tif

dataset/.gitignore

@@ -1 +1,4 @@
 *.csv
+RaCA/*
+RaCA_SimCLR_all/*
+RaCA_SimCLR/*