Update README.md with new model card content

README.md

@@ -13,6 +13,22 @@ pipeline_tag: text-generation
 
 Falcon-RW-1B is a 1B parameters causal decoder-only model built by [TII](https://www.tii.ae/) and trained on 350B tokens of [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb). The architecture of the model is adopted from the GPT-3 paper ([Brown et al., 2020](https://arxiv.org/abs/2005.14165)) but it uses ALiBi.
 
+## Links
+
+* [Falcon Quickstart Notebook](https://www.kaggle.com/code/laxmareddypatlolla/falcon-quickstart-notebook)
+* [Falcon API Documentation](https://keras.io/keras_hub/api/models/falcon/)
+* [Falcon Model Card](https://huggingface.co/docs/transformers/en/model_doc/falcon)
+* [KerasHub Beginner Guide](https://keras.io/guides/keras_hub/getting_started/)
+* [KerasHub Model Publishing Guide](https://keras.io/guides/keras_hub/upload/)
+
+## Presets
+
+The following model checkpoints are provided by the Keras team. Full code examples for each are available below.
+| Preset name    | Parameters | Description                                      |
+|----------------|------------|--------------------------------------------------|
+| falcon_refinedweb_1b_en |   1.31B  | 24-layer Falcon model (Falcon with 1B parameters), trained on 350B tokens of RefinedWeb dataset.|
+
+
 ## Use
 
 ### Direct Use
@@ -82,3 +98,53 @@ The architecture is adapted from the GPT-3 paper ([Brown et al., 2020](https://a
 }
 ```
 
+## Example Usage
+```Python
+
+import os
+
+os.environ["KERAS_BACKEND"] = "jax"
+
+import keras
+import keras_hub
+
+# When running only inference, bfloat16 saves memory usage significantly.
+keras.config.set_floatx("bfloat16")
+
+causal_lm = keras_hub.models.FalconCausalLM.from_preset(
+    "falcon_refinedweb_1b_en"
+)
+causal_lm.summary()
+
+outputs = causal_lm.generate([
+    "What is Jax?",
+    "Give me your best brownie recipe.",
+], max_length=512)
+
+```
+
+## Example Usage with Hugging Face URI
+
+```Python
+
+import os
+
+os.environ["KERAS_BACKEND"] = "jax"
+
+import keras
+import keras_hub
+
+# When running only inference, bfloat16 saves memory usage significantly.
+keras.config.set_floatx("bfloat16")
+
+causal_lm = keras_hub.models.FalconCausalLM.from_preset(
+    "hf://keras/falcon_refinedweb_1b_en"
+)
+causal_lm.summary()
+
+outputs = causal_lm.generate([
+    "What is Jax?",
+    "Give me your best brownie recipe.",
+], max_length=512)
+
+```