Update README.md

README.md

@@ -8,16 +8,17 @@ language:
 - su
 license: llama3
 ---
-# Llama3 8B CPT Sahabat AI v1 Instruct
+# Llama3 8B CPT Sahabat-AI v1 Instruct
 
-Sahabat AI  is a collection of Large Language Models (LLMs) which has been pretrained and instruct-tuned for Indonesian languages.
+**Sahabat-AI** (Indonesian language for “close friends”) is a collection of Large Language Models (LLMs) which has been pretrained and instruct-tuned for Indonesian language and its various dialects. Sahabat-AI ecosystem is co-initiated by Indonesian tech and telecommunication companies:  GoTo Group and Indosat Ooredoo Hutchison.
 
-Llama3 8B CPT Sahabat AI v1 Instruct is an Indonesian-focused model which has been fine-tuned with around **448,000 Indonesian instruction-completion pairs** alongside an Indonesian-dialect pool consisting of **96,000 instruction-completion pairs in Javanese** and **98,000 instruction-completion pairs in Sundanese**. Additionally, we added a pool of **129,000 instruction-completion pairs in English**.
+Llama3 8B CPT Sahabat-AI v1 Instruct is an Indonesian-focused model which has been fine-tuned with around **448,000 Indonesian instruction-completion pairs** alongside an Indonesian-dialect pool consisting of **96,000 instruction-completion pairs in Javanese** and **98,000 instruction-completion pairs in Sundanese**. Additionally, we added a pool of **129,000 instruction-completion pairs in English**.
 
 Sahabat is Indonesian for "Close Friends."
 
 - **Developed by:** PT GoTo Gojek Tokopedia Tbk,  AI Singapore
-- **Funded by:** PT GoTo Gojek Tokopedia Tbk, AI Singapore  
+- **Funded by:** PT GoTo Gojek Tokopedia Tbk, AI Singapore
+- **Supported by:**  PT Indosat Ooredoo Hutchison
 - **Model type:** Decoder
 - **Languages:** English, Indonesian, Javanese, Sundanese
 - **License:** [Llama3 Community License](https://huggingface.co/meta-llama/Meta-Llama-3-8B/blob/main/LICENSE)
@@ -25,12 +26,12 @@ Sahabat is Indonesian for "Close Friends."
 ## Model Details
 
 ### Model Description
-We performed instruction tuning in Indonesian, Javanese, Sundanese as well as English on our [continued pre-trained Llama3 8B CPT Sahabat AI v1 base](https://huggingface.co/GoToCompany/llama3-8b-cpt-sahabatai-v1-base), a decoder model using the Llama3 architecture, to create Llama3 8B CPT Sahabat AI v1 Instruct.
+We performed instruction tuning in Indonesian, Javanese, Sundanese as well as English on our [continued pre-trained Llama3 8B CPT Sahabat-AI v1 base](https://huggingface.co/GoToCompany/llama3-8b-cpt-sahabatai-v1-base), a decoder model using the Llama3 architecture, to create Llama3 8B CPT Sahabat-AI v1 Instruct.
 
 For tokenisation, the model employs the default tokenizer used in Llama-3-8B. The model has a context length of 8192.
 
 ### Benchmark Performance
-We evaluated Llama3 8B CPT Sahabat AI V1 Instruct on both general language capabilities and instruction-following capabilities.
+We evaluated Llama3 8B CPT Sahabat-AI V1 Instruct on both general language capabilities and instruction-following capabilities.
 
 #### General Language Capabilities
 For the evaluation of general language capabilities, we employed the 
@@ -49,7 +50,7 @@ The evaluation was done **zero-shot** with native prompts on a sample of 100-100
 
 
 #### Instruction-following Capabilities
-Since Llama3 8B CPT Sahabat AI v1 Instruct is an instruction-following model, we also evaluated it on instruction-following capabilities with the [IFEval](https://arxiv.org/abs/2311.07911) dataset.
+Since Llama3 8B CPT Sahabat-AI v1 Instruct is an instruction-following model, we also evaluated it on instruction-following capabilities with the [IFEval](https://arxiv.org/abs/2311.07911) dataset.
 
   As this dataset was in English, the linguists and native speakers in the team worked together to filter, localize and translate the dataset into the respective target languages to ensure that the examples remained reasonable, meaningful and natural.
 
@@ -173,7 +174,7 @@ IFEval evaluates a model's ability to adhere to constraints provided in the prom
 </table>
 
 
-Llama3 8B CPT Sahabat AI v1 Instruct can be run using the 🤗 Transformers library 
+Llama3 8B CPT Sahabat-AI v1 Instruct can be run using the 🤗 Transformers library 
 ```python
 # Please use transformers==4.45.0
 
@@ -232,15 +233,31 @@ Current Sahabat models, including this commercially permissive release, have not
 
 ## Technical Specifications
 ### Fine-Tuning Details
-Llama3 8B CPT Sahabat AI v1 Instruct was built using a combination of a full parameter fine-tune, on-policy alignment, and model merges of the best performing checkpoints. The training process for fine-tuning was approximately 4 hours, with alignment taking 2 hours, both on 8x H100-80GB GPUs.
+Llama3 8B CPT Sahabat-AI v1 Instruct was built using a combination of a full parameter fine-tune, on-policy alignment, and model merges of the best performing checkpoints. The training process for fine-tuning was approximately 4 hours, with alignment taking 2 hours, both on 8x H100-80GB GPUs.
 
 ## Data
-Llama3 8B CPT Sahabat AI v1 Instruct was trained on a wide range of synthetic instructions, alongside publicly available instructions hand-curated by the team with the assistance of native speakers. In addition, special care was taken to ensure that the datasets used had commercially permissive licenses through verification with the original data source. 
+Llama3 8B CPT Sahabat-AI v1 Instruct was trained on a wide range of synthetic instructions, alongside publicly available instructions hand-curated by the team with the assistance of native speakers. In addition, special care was taken to ensure that the datasets used had commercially permissive licenses through verification with the original data source. 
 
-## Call for Contributions
-We encourage researchers, developers, and language enthusiasts to actively contribute to the enhancement and expansion of Sahabat. Contributions can involve identifying and reporting bugs, sharing pre-training, instruction, and preference data, improving documentation usability, proposing and implementing new model evaluation tasks and metrics, or training versions of the model in additional Indonesian languages. Join us in shaping the future of Sahabat by sharing your expertise and insights to make these models more accessible, accurate, and versatile.
+## Call for Collaboration 
+Sahabat-AI (Indonesian language for “close friends”) a **local open source Large Language Model (LLM) ecosystem in Indonesian language**, co-initiated by Indonesian tech and telecommunication companies:  GoTo Group and Indosat Ooredoo Hutchison. 
+Sahabat-AI ecosystem aims to empower Indonesians who want to develop AI-based services and applications using Bahasa Indonesia and its various local dialects. 
 
-## The Team (by ascending alphabetical order)
+We are supported by research centers and global tech experts such as AI Singapore and Tech Mahendra to train the model to gain general language understanding.
+
+We also have collaborated with key top Indonesia universities such as University of Indonesia, Gadjah Mada University, Bogor Institute of Agriculture, Bandung Institute of Technology, including top Indonesia media groups, such as Kompas Media Group and Republika to train and enrich the model in Bahasa Indonesia, ensuring optimum provision of local context and cultural relevance.
+
+We would like to invite **researchers, developers, and language enthusiasts** to actively contribute to the enhancement and expansion of Sahabat-AI. 
+Your collaborations can involve:
+- Identifying and reporting technical issues
+- Sharing pre-training, instruction, and preference data
+- Improving documentation usability
+- Proposing and implementing new model evaluation tasks and metrics 
+
+Join us in shaping the future of Sahabat-AI by sharing your expertise and insights to make these models more accessible, accurate, and versatile.
+
+You can contribute your ideas through [this form.](https://docs.google.com/forms/d/1_us969eQtEooYOn4XkvGkdP5VHOyCbO6L_sd9kTMnaA/edit)
+
+## The Development Team (in ascending alphabetical order)
 
 ### AI Singapore
 Chan Adwin<br>