Update README.md

README.md

@@ -32,7 +32,7 @@ It is designed to **push the boundaries** of open-source agentic LLMs, excelling
 - **License:** Apache 2.0  
 - **Architecture:** Meta-Llama 3.1-405B Instruct
 - **Training Data:** CALM-IT
-- **Fine-tuning Framework:** Oumi  
+- **Fine-tuning Framework:** [Oumi](https://github.com/oumi-ai/oumi)  
 - **Training Hardware:** 8 NVIDIA H100 GPUs  
 - **Training Duration:** ~6.5 days  
 - **Evaluation Benchmarks:** MultiWOZ 2.4, BFCL V3, API-Bank  
@@ -75,7 +75,7 @@ TODO: Add BFCL results
 ## 💡 How to Use CALM-405B
 🚨 It requires 16xH100 NVIDIA GPUs for Inference.
 
-### 🏗 How to Load the Model
+### 🏗 How to Load the Model using HuggingFace
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
@@ -83,12 +83,16 @@ tokenizer = AutoTokenizer.from_pretrained("uiuc-convai/CALM-8B")
 model = AutoModelForCausalLM.from_pretrained("uiuc-convai/CALM-8B")
 ```
 
-<!-- TODO -->
-### 🛠 Example Inference
-```python
-TODO
-```
+### 🛠 Example Oumi Inference
+CALM-405B likely requires multi-node inference as most single nodes support up to 640GB of GPU VRAM. To run multi-node inference, we recommend [vLLM](https://docs.vllm.ai/en/latest/serving/distributed_serving.html)
 
+### 🛠 Example Oumi Fine-Tuning
+```bash
+pip install oumi
+
+# See oumi_train.yaml in this model's /oumi/ directory.
+oumi train -c ./oumi_train.yaml
+```
 
 More fine-tuning and **community-driven** optimizations are planned to enhance real-world usability.