Update README.md

README.md

@@ -3,23 +3,28 @@ base_model: meta-llama/Meta-Llama-3-8B-Instruct
 library_name: peft
 ---
 
-# MISHANM/Sindhi_text_generation_Llama3_8B_instruction
-
-This model is fine-tuned for the Sindhi language, capable of answering queries and translating text Between English and Sindhi . It leverages advanced natural language processing techniques to provide accurate and context-aware responses.
+# MISHANM/Multilingual_Llama-3-8B-Instruct
 
+This model is fine-tuned for Multi languages , capable of answering queries and translating text from English to Multiple languages . It leverages advanced natural language processing techniques to provide accurate and context-aware responses.
 
 
 ## Model Details
-1. Language: Sindhi 
-2. Tasks: Question Answering, Translation (English to Sindhi )
-3. Base Model: meta-llama/Meta-Llama-3-8B-Instruct
+This model is based on meta-llama/Llama-3.2-3B-Instruct and has been LoRA finetuned on Indic datasets:
+1.Gujarati
+2.Kannada
+3.Hindi
+4.Odia
+5.Punjabi
+6.Bengali
+7.Tamil
+8.Telugu 
 
 
 
 # Training Details
 
-The model is trained on approx 29K instruction samples.
-1. GPUs: 4*AMD Radeon™ PRO V620
+The model is trained on approx 321K instruction samples.
+1. GPUs: 2*AMD Instinct™ MI210 Accelerators
   
    
 
@@ -34,7 +39,7 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
 # Load the fine-tuned model and tokenizer
-model_path = "MISHANM/Sindhi_text_generation_Llama3_8B_instruction"
+model_path = "MISHANM/Multilingual_Llama-3-8B-Instruct"
 model = AutoModelForCausalLM.from_pretrained(model_path)
 
 # Wrap the model with DataParallel if multiple GPUs are available
@@ -53,7 +58,7 @@ def generate_text(prompt, max_length=1000, temperature=0.9):
     messages = [
         {
             "role": "system",
-            "content": "You are a Sindhi language expert and linguist, with same knowledge give answers in Sindhi language. ",
+            "content": "You are a language expert and linguist, with same knowledge give response in ().", #In place of "()" write your desired language in which response is required. ",
         },
         {"role": "user", "content": prompt}
     ]
@@ -69,7 +74,7 @@ def generate_text(prompt, max_length=1000, temperature=0.9):
     return tokenizer.decode(output[0], skip_special_tokens=True)
 
 # Example usage
-prompt = """Write a poem LLM ."""
+prompt = """Write a story about LLM ."""
 translated_text = generate_text(prompt)
 print(translated_text)
 
@@ -78,9 +83,9 @@ print(translated_text)
 
 ## Citation Information
 ```
-@misc{MISHANM/Sindhi_text_generation_Llama3_8B_instruction,
+@misc{MISHANM/Multilingual_Llama-3-8B-Instruct,
   author = {Mishan Maurya},
-  title = {Introducing Fine Tuned LLM for Sindhi Language},
+  title = {Introducing Fine Tuned LLM for Indic Languages},
   year = {2024},
   publisher = {Hugging Face},
   journal = {Hugging Face repository},