diff --git a/README.md b/README.md
index e5e5b97..9286445 100644
--- a/README.md
+++ b/README.md
@@ -87,6 +87,9 @@ geo_inference -i <image> -br <bands_requested> -m <model> -wd <work_dir> -ps <pa
 - `-cls`, `--classes`: The number of classes that model outputs, Default = 5
 - `-mg`, `--mgpu`: Whether to use multi-gpu processing or not, Default = False
 - `-pr`, `--prediction_thr` : Prediction probability Threshold (fraction of 1) to use. Default = 0.3
+- `-tr`, `--transformers`: Allow Test-time augmentations.  
+- `tr_f`, `transformer_flip`: Perform horizontal and vertical flips.  
+- `tr_e`, `transformer_rotate`: perform 90 degree rotation.  
 
 
 You can also use the `-h` option to get a list of supported arguments:
@@ -110,7 +113,10 @@ geo_inference = GeoInference(
     multi_gpu=False,
     gpu_id=0, 
     num_classes=5,
-    prediction_threshold=0.3
+    prediction_threshold=0.3,
+    transformers=True,
+    transformer_flip=False,
+    transformer_rotate=True,
 )
 
 # Perform feature extraction on a TIFF image
@@ -143,6 +149,9 @@ Initiating the `GeoInference` class takes the following parameters:
 - `gpu_id`: The ID of the GPU to use for feature extraction. Default is `0`.
 - `num_classes`: The number of classes that the TorchScript model outputs. Default is `5`.
 - `prediction_threshold`: Prediction probability Threshold (fraction of 1) to use. Default is `0.3`.  
+- `transformers`: Allow Test-time augmentations.  
+- `transformer_flip`: Perform horizontal and vertical flips.  
+- `transformer_rotate`: perform 90 degree rotation.  
 
 Calling the GeoInference object takes the following parameters:  
 - `inference_input`: Path to Geotiff. 
diff --git a/VERSION b/VERSION
index 6a0ca2d..56fea8a 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-2.0.7
\ No newline at end of file
+3.0.0
\ No newline at end of file
diff --git a/geo_inference/__init__.py b/geo_inference/__init__.py
index 7bcd38b..571b01f 100644
--- a/geo_inference/__init__.py
+++ b/geo_inference/__init__.py
@@ -1,4 +1,4 @@
 """Geo-inference: Extract features from high-resolution geospatial imagery using foundation models"""
 
 __author__ = "Victor Alhassan"
-__version__ = "2.0.7"
+__version__ = "3.0.0"
diff --git a/geo_inference/geo_inference.py b/geo_inference/geo_inference.py
index d849e5a..0671a3b 100644
--- a/geo_inference/geo_inference.py
+++ b/geo_inference/geo_inference.py
@@ -59,6 +59,9 @@ class GeoInference:
         gpu_id (int): The ID of the GPU to use for inference (if device is "gpu").
         num_classes (int) : The number of classes in the output of the model.
         prediction_threshold (float): Prediction probability Threshold (fraction of 1) to use.
+        transformers (bool): Allow Test-time augmentations.
+        transformer_flip (bool): Perform horizontal and vertical flips.
+        transformer_rotate (bool): Perform 90 degree rotation.
 
     Attributes:
         work_dir (Path): The directory where the model and output files will be saved.
diff --git a/pyproject.toml b/pyproject.toml
index 46a2e61..4ad8360 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -37,7 +37,7 @@ geo_inference = "geo_inference.geo_inference:main"
 include = ["geo_inference*"]
 
 [tool.bumpver]
-current_version = "2.0.7"
+current_version = "3.0.0"
 version_pattern = "MAJOR.MINOR.PATCH"
 commit_message = "bump version {old_version} -> {new_version}"
 commit = true