Update readme (#9)

Author: james77777778

README.md

@@ -1,14 +1,123 @@
 # Keras Image Models
 
-## Unit Tests
+## Description
+
+**K**eras **Im**age **M**odels (`kimm`) is a collection of image models, blocks and layers written in Keras 3. The goal is to offer SOTA models with pretrained weights in a user-friendly manner.
+
+## Installation
 
 ```bash
-# KERAS_BACKEND=jax|numpy|tensorflow|torch
-CUDA_VISIBLE_DEVICES= KERAS_BACKEND=tensorflow pytest
+# In a working [jax/tensorflow/torch/numpy] backend environment
+pip install keras kimm
+```
+
+## Quickstart
+
+### Use Pretrained Model
+
+```python
+from keras import random
+
+import kimm
+
+# Use `kimm.list_models` to get the list of available models
+print(kimm.list_models())
+
+# Specify the name and other arguments to filter the result
+print(kimm.list_models("efficientnet", has_pretrained=True))  # fuzzy search
+
+# Initialize the model with pretrained weights
+model = kimm.models.EfficientNetV2B0(weights="imagenet")
+
+# Predict
+x = random.uniform([1, 192, 192, 3]) * 255.0
+y = model.predict(x)
+print(y.shape)
+
+# Initialize the model as a feature extractor with pretrained weights
+model = kimm.models.EfficientNetV2B0(
+    as_feature_extractor=True, weights="imagenet"
+)
+
+# Extract features for downstream tasks
+y = model.predict(x)
+print(y.keys())
+print(y["BLOCK5_S32"].shape)
+```
+
+### Transfer Learning
+
+```python
+from keras import layers
+from keras import models
+from keras import random
+
+import kimm
+
+# Initialize the model as a backbone with pretrained weights
+backbone = kimm.models.EfficientNetV2B0(
+    input_shape=[224, 224, 3],
+    include_top=False,
+    pooling="avg",
+    weights="imagenet",
+)
+
+# Freeze the backbone for transfer learning
+backbone.trainable = False
+
+# Construct the model with new head
+inputs = layers.Input([224, 224, 3])
+x = backbone(inputs, training=False)
+x = layers.Dropout(0.2)(x)
+outputs = layers.Dense(2)(x)
+model = models.Model(inputs, outputs)
+
+# Train the new model (put your own logic here)
+
+# Predict
+x = random.uniform([1, 224, 224, 3]) * 255.0
+y = model.predict(x)
+print(y.shape)
 ```
 
-## Work in Progress
+## License
+
+Please refer to [timm](https://github.com/huggingface/pytorch-image-models#licenses) as this project is built upon it.
+
+### `kimm` Code
+
+The code here is licensed Apache 2.0.
+
+## Acknowledgements
 
-- Test pretrained weights
+Thanks for these awesome projects that were used in `kimm`
 
-## Acknowledgments
+- [https://github.com/keras-team/keras](https://github.com/keras-team/keras)
+- [https://github.com/huggingface/pytorch-image-models](https://github.com/huggingface/pytorch-image-models)
+
+## Citing
+
+### BibTeX
+
+```bash
+@misc{rw2019timm,
+  author = {Ross Wightman},
+  title = {PyTorch Image Models},
+  year = {2019},
+  publisher = {GitHub},
+  journal = {GitHub repository},
+  doi = {10.5281/zenodo.4414861},
+  howpublished = {\url{https://github.com/rwightman/pytorch-image-models}}
+}
+```
+
+```bash
+@misc{hy2024kimm,
+  author = {Hongyu Chiu},
+  title = {Keras Image Models},
+  year = {2024},
+  publisher = {GitHub},
+  journal = {GitHub repository},
+  howpublished = {\url{https://github.com/james77777778/kimm}}
+}
+```

kimm/utils/model_registry.py

@@ -9,6 +9,23 @@
 MODEL_REGISTRY: typing.List[typing.Dict[str, typing.Union[str, bool]]] = []
 
 
+def _match_string(query: str, target: str):
+    query = query.lower().replace(" ", "").replace("_", "").replace(".", "")
+    target = target.lower()
+    matched_idx = -1
+    for q_char in query:
+        matched = False
+        for idx, t_char in enumerate(target):
+            if matched:
+                break
+            if q_char == t_char and idx > matched_idx:
+                matched_idx = idx
+                matched = True
+        if not matched:
+            return False
+    return True
+
+
 def clear_registry():
     MODEL_REGISTRY.clear()
 
@@ -47,9 +64,10 @@ def list_models(
         result_names.add(info["name"])
         need_remove = False
 
-        # filter by the args
-        if name is not None and name.lower() not in info["name"].lower():
-            need_remove = True
+        # match string (simple implementation)
+        if name is not None:
+            need_remove = not _match_string(name, info["name"])
+        # filter by support_feature and has_pretrained
         if (
             support_feature is not None
             and info["support_feature"] is not support_feature