Update README.md

README.md

@@ -7,16 +7,14 @@ base_model:
 pipeline_tag: text-generation
 ---
 
-# Bayon Instruct Preview
+# Bayon Instruct
 
-**Bayon Instruct Preview** is a fine-tune of the first native, from-scratch decoder-only generative Khmer language model [Bayon](https://huggingface.co/attentionlab/bayon). Bayon Instruct Preview <!--is the world's first autoregressive decoder-only model to feature chain of thought (CoT) reasoning in the Khmer language. This model--> was created and evaluated in our AACL Paper, you can learn more about the model and its evaluation below. 
+**Bayon Instruct** is a fine-tune of the first native, from-scratch decoder-only generative Khmer language model [Bayon](https://huggingface.co/attentionlab/bayon). Bayon Instruct was created and evaluated in our AACL Paper, you can learn more about the model and its evaluation below. 
 
 ---
 
 ## 🚀 Key Features
 
-<!--* **Reasoning:** This model is capable of generating responses with minimal thinking or not at all. 
-* **Tool Use Capabilities:** Bayon was trained to support basic retrieval augmented generation (RAG) along with math tools. -->
 * **Minimal Language Bleeding:** Even if not spoken to in Khmer, Bayon will almost **never** generate a response in another language other than Khmer, see our paper below for further information.
 
 ---
@@ -30,7 +28,7 @@ import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
 # 1. Specify the model repository path
-model_id = "attentionlab/bayon-it-preview"
+model_id = "attentionlab/bayon-it"
 
 # 2. Load the custom tokenizer and optimized model weights
 tokenizer = AutoTokenizer.from_pretrained(model_id)