update Readme with Mixins (#70)

* update readme with Mixins
* Apply suggestions from code review

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Aniket Maurya <[email protected]>

Author: 3 people

README.md

@@ -161,3 +161,71 @@ trainer.fit(
 ```
 
 </details>
+
+## Model Registry Mixins
+
+Lightning Models provides mixin classes that simplify pushing models to and pulling models from the registry.
+These mixins can be integrated directly into the model classes.
+
+### Available Mixins
+
+1. **PickleRegistryMixin**: For serializing any Python class with pickle
+2. **PyTorchRegistryMixin**: For PyTorch models, preserving both weights and constructor arguments
+
+Using these mixins provides several advantages:
+
+- Direct integration into the model classes
+- Simplified save/load workflow
+- Automatic handling of model metadata and constructor arguments
+- Version management support
+
+### Using `PickleRegistryMixin`
+
+Add the mixin to a Python class for seamless registry integration:
+
+```python
+from litmodels.integrations.mixins import PickleRegistryMixin
+
+
+class MyModel(PickleRegistryMixin):
+    def __init__(self, param1, param2):
+        self.param1 = param1
+        self.param2 = param2
+        # Your model initialization code
+
+
+# Create and push a model instance
+model = MyModel(param1=42, param2="hello")
+model.push_to_registry(name="my-org/my-team/my-model")
+
+# Later, pull the model
+loaded_model = MyModel.pull_from_registry(name="my-org/my-team/my-model")
+```
+
+### Using `PyTorchRegistryMixin`
+
+This mixin preserves both the model architecture and weights:
+
+```python
+import torch
+from litmodels.integrations.mixins import PyTorchRegistryMixin
+
+
+# Important: PyTorchRegistryMixin must be first in the inheritance order
+class MyTorchModel(PyTorchRegistryMixin, torch.nn.Module):
+    def __init__(self, input_size, hidden_size=128):
+        super().__init__()
+        self.linear = torch.nn.Linear(input_size, hidden_size)
+        self.activation = torch.nn.ReLU()
+
+    def forward(self, x):
+        return self.activation(self.linear(x))
+
+
+# Create and push the model
+model = MyTorchModel(input_size=784)
+model.push_to_registry(name="my-org/my-team/torch-model")
+
+# Pull the model with the same architecture
+loaded_model = MyTorchModel.pull_from_registry(name="my-org/my-team/torch-model")
+```