Update README.md

README.md

@@ -1,3 +1,244 @@
----
-license: apache-2.0
----
+---
+tags:
+- text-generation-inference
+- text-generation
+- Sentiment Analysis
+- qlora
+- peft
+license: apache-2.0
+library_name: transformers
+widget:
+- messages:
+  - role: user
+    content: What is your name?
+language:
+- en
+- ro
+pipeline_tag: text-generation
+model-index:
+- name: CognitivessAI/cognitivess
+  results:
+  - task:
+      type: text-generation
+      name: Text Generation
+    metrics:
+      - name: Perplexity
+        type: perplexity
+        value: 7.5  # Replace with your actual perplexity value
+      - name: ROUGE-L
+        type: rouge-l
+        value: 0.85  # Replace with your actual ROUGE-L score
+base_model: CognitivessAI/bella-2-8b
+model_type: CognitivessForCausalLM
+quantization_config:
+  load_in_8bit: true
+  llm_int8_threshold: 6.0
+fine_tuning:
+  method: qlora
+  peft_type: LORA
+inference:
+  parameters:
+    max_new_tokens: 8192
+    temperature: 0.7
+    top_p: 0.95
+    do_sample: true
+---
+
+<div align="center">
+ <img src="https://cdn-uploads.huggingface.co/production/uploads/65ec00afa735404e87e1359e/u5qyAgn_2-Bh46nzOFlcI.png">
+ <h2>Accessible and portable generative AI solutions for developers and businesses.</h2>
+ </div>
+
+ <p align="center" style="margin-top: 0px;">
+     <a href="https://cognitivess.com">
+    <span class="link-text" style=" margin-right: 5px;">Website</span>
+  </a> |
+  <a href="https://bella.cognitivess.com">
+    <span class="link-text" style=" margin-right: 5px;">Demo</span>
+  </a> |
+  <a href="https://github.com/Cognitivess/cognitivess">
+    <img src="https://github.githubassets.com/assets/GitHub-Mark-ea2971cee799.png" alt="GitHub Logo" style="width:20px; vertical-align: middle; display: inline-block; margin-right: 5px; margin-left: 5px; margin-top: 0px; margin-bottom: 0px;"/>
+    <span class="link-text" style=" margin-right: 5px;">GitHub</span>
+  </a>
+</p>
+
+# Cognitivess
+
+Cognitivess is an advanced language model developed by Cognitivess AI, based in Bucharest, Romania. This model, fine-tuned from the Bella-2-8b base, utilizes Quantized Low-Rank Adaptation (QLoRA) techniques to deliver high-quality text generation while maintaining efficiency.
+
+Key features:
+- Built on the LLaMA architecture
+- Fine-tuned using QLoRA for optimal performance and resource utilization
+- Capable of generating text in both English and Romanian
+- Specialized in tasks such as text generation, sentiment analysis, and general question-answering
+- Designed to provide clear, concise, and informative responses in a conversational manner
+
+Cognitivess aims to serve as a versatile AI assistant, capable of handling a wide range of queries and tasks while maintaining a friendly and professional demeanor. Whether you need help with analysis, creative writing, or just engaging in informative dialogue, Cognitivess is equipped to assist.
+
+This model represents Cognitivess AI's commitment to advancing natural language processing technology and making it accessible for various applications.
+
+
+***Under the Cognitivess Open Model License, Cognitivess AI confirms:***
+- Models are commercially usable. 
+- You are free to create and distribute Derivative Models. 
+- Cognitivess does not claim ownership to any outputs generated using the Models or Derivative Models.
+
+### Intended use
+
+Cognitivess is a multilingual chat model designed to support a variety of languages including English, Romanian, Spanish, French, German, and many more, intended for diverse language applications.
+
+
+**Model Developer:** Cognitivess AI
+
+**Model Dates:** Cognitivess was trained between July 2024.
+
+**Data Freshness:** The pretraining data has a cutoff of June 2024. Training will continue beyond the current data cutoff date to incorporate new data as it becomes available.
+
+
+### Model Architecture:
+
+Cognitivess model architecture is Transformer-based and trained with a sequence length of 8192 tokens.
+
+**Architecture Type:** Transformer (auto-regressive language model)
+
+
+
+Try this model on [bella.cognitivess.com](https://bella.cognitivess.com/) now.
+
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/65ec00afa735404e87e1359e/CQeAV4lwbQp1G8H5n4uWx.png)
+
+
+# Usage
+
+```python
+
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from peft import PeftModel, PeftConfig
+
+# Set the device
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+print(f"Using device: {device}")
+
+# Load the tokenizer
+tokenizer = AutoTokenizer.from_pretrained("CognitivessAI/cognitivess")
+
+# Load the PEFT configuration
+peft_config = PeftConfig.from_pretrained("CognitivessAI/cognitivess")
+
+# Load the base model
+base_model = AutoModelForCausalLM.from_pretrained(
+    peft_config.base_model_name_or_path,
+    device_map="auto",
+    torch_dtype=torch.float16
+)
+
+# Load the PEFT model
+model = PeftModel.from_pretrained(base_model, "CognitivessAI/cognitivess")
+
+# Move the model to the appropriate device
+model = model.to(device)
+
+# Set the model to evaluation mode
+model.eval()
+
+# Function for text generation using the chat template
+def generate_text(model, tokenizer, input_text, max_length=8192, temperature=0.7, top_p=0.95):
+    messages = [
+        {"role": "user", "content": input_text}
+    ]
+    chat_input = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+    inputs = tokenizer(chat_input, return_tensors='pt', padding=True, truncation=True, max_length=8192)
+    input_ids = inputs['input_ids'].to(device)
+    attention_mask = inputs['attention_mask'].to(device)
+    try:
+        generated_text_ids = model.generate(
+            input_ids,
+            attention_mask=attention_mask,
+            max_length=max_length,
+            temperature=temperature,
+            top_p=top_p,
+            do_sample=True,
+            eos_token_id=tokenizer.eos_token_id
+        )
+        generated_text = tokenizer.decode(generated_text_ids[0], skip_special_tokens=True)
+        # Extract the assistant's response
+        response = generated_text.split("GPT4 Correct Assistant")[-1].strip()
+        return response
+    except Exception as e:
+        print(f"Error in text generation: {e}")
+        return "I'm sorry, I encountered an error while generating a response."
+
+# Test the model
+test_prompt = "Who are you?"
+generated_response = generate_text(model, tokenizer, test_prompt, max_length=100)
+print(f"Generated response:\n{generated_response}")
+
+print("Testing completed.")import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from peft import PeftModel, PeftConfig
+
+# Set the device
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+print(f"Using device: {device}")
+
+# Load the tokenizer
+tokenizer = AutoTokenizer.from_pretrained("CognitivessAI/cognitivess")
+
+# Load the PEFT configuration
+peft_config = PeftConfig.from_pretrained("CognitivessAI/cognitivess")
+
+# Load the base model
+base_model = AutoModelForCausalLM.from_pretrained(
+    peft_config.base_model_name_or_path,
+    device_map="auto",
+    torch_dtype=torch.float16
+)
+
+# Load the PEFT model
+model = PeftModel.from_pretrained(base_model, "CognitivessAI/cognitivess")
+
+# Move the model to the appropriate device
+model = model.to(device)
+
+# Set the model to evaluation mode
+model.eval()
+
+# Function for text generation using the chat template
+def generate_text(model, tokenizer, input_text, max_length=8192, temperature=0.7, top_p=0.95):
+    messages = [
+        {"role": "user", "content": input_text}
+    ]
+    chat_input = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+    inputs = tokenizer(chat_input, return_tensors='pt', padding=True, truncation=True, max_length=8192)
+    input_ids = inputs['input_ids'].to(device)
+    attention_mask = inputs['attention_mask'].to(device)
+    try:
+        generated_text_ids = model.generate(
+            input_ids,
+            attention_mask=attention_mask,
+            max_length=max_length,
+            temperature=temperature,
+            top_p=top_p,
+            do_sample=True,
+            eos_token_id=tokenizer.eos_token_id
+        )
+        generated_text = tokenizer.decode(generated_text_ids[0], skip_special_tokens=True)
+        # Extract the assistant's response
+        response = generated_text.split("GPT4 Correct Assistant")[-1].strip()
+        return response
+    except Exception as e:
+        print(f"Error in text generation: {e}")
+        return "I'm sorry, I encountered an error while generating a response."
+
+# Test the model
+test_prompt = "Who are you?"
+generated_response = generate_text(model, tokenizer, test_prompt, max_length=100)
+print(f"Generated response:\n{generated_response}")
+
+
+```
+
+**Contact:**
+<a href="mailto:hello@cognitivess.com">hello@cognitivess.com</a>