update

README.md

@@ -10,4 +10,66 @@ pinned: false
 short_description: Chat with AI with ⚡Lightning Speed
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# H GO
+
+Inspired by Google Go, H GO is a concise and efficient chat interface that leverages various language models hosted on Hugging Face. It uses Gradio for the user interface and supports different models to cater to diverse needs.
+
+## Features
+
+- **Multiple Model Support**: Choose from various models such as Nous Hermes, StarChat, Mistral, and Phi.
+- **Real-time Interaction**: Get quick and concise responses from the selected model.
+- **Customizable**: Easily switch models to suit your specific requirements.
+
+## Setup
+
+### Prerequisites
+
+- Python 3.7+
+- Gradio
+- Hugging Face Hub
+- Git LFS (Large File Storage)
+
+### Installation
+
+1. Clone the repository:
+
+ ```bash
+ # Make sure you have git-lfs installed (https://git-lfs.com)
+ git lfs install
+
+ # When prompted for a password, use an access token with write permissions.
+ # Generate one from your settings: https://huggingface.co/settings/tokens
+ git clone https://huggingface.co/spaces/HarshanaLF/Real-Time-Chat-with-AI
+
+ # If you want to clone without large files - just their pointers
+ GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/spaces/HarshanaLF/Real-Time-Chat-with-AI
+ ```
+
+2. Navigate to the project directory:
+
+ ```bash
+ cd Real-Time-Chat-with-AI
+ ```
+
+3. Install the required dependencies:
+
+ ```bash
+ pip install gradio huggingface_hub
+ ```
+
+## Usage
+
+Run the application using the following command:
+
+```bash
+python app.py
+
+
+## Model Descriptions
+
+- **Nous Hermes Mixtral 8x7B DPO**: A robust model designed for detailed and nuanced conversation handling.
+- **StarChat2 15b**: A large-scale model optimized for general chat interactions with a wide range of topics.
+- **Mistral 7B v0.3**: A smaller, efficient model suitable for fast and responsive chat applications.
+- **Phi 3 mini**: A compact model focusing on instructive and concise responses.
+- **Mixtral 8x7B**: A versatile model capable of handling various conversational contexts effectively.
+```

app.py

@@ -1,44 +1,52 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
 
-
+# Function to return the appropriate client based on the model selected
 def client_fn(model):
- if "Nous" in model:
- return InferenceClient("NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO")
- elif "Star" in model:
- return InferenceClient("HuggingFaceH4/starchat2-15b-v0.1")
- elif "Mistral" in model:
- return InferenceClient("mistralai/Mistral-7B-Instruct-v0.3")
- elif "Phi" in model:
- return InferenceClient("microsoft/Phi-3-mini-4k-instruct")
- else: 
- return InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
+ model_map = {
+ "Nous Hermes Mixtral 8x7B DPO": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
+ "StarChat2 15b": "HuggingFaceH4/starchat2-15b-v0.1",
+ "Mistral 7B v0.3": "mistralai/Mistral-7B-Instruct-v0.3",
+ "Phi 3 mini": "microsoft/Phi-3-mini-4k-instruct",
+ "Mixtral 8x7B": "mistralai/Mixtral-8x7B-Instruct-v0.1"
+ }
+ return InferenceClient(model_map.get(model, "mistralai/Mixtral-8x7B-Instruct-v0.1"))
 
-system_instructions1 = "[SYSTEM] Your task is to Answer the question. Keep conversation very short, clear and concise. The expectation is that you will avoid introductions and start answering the query directly, Only answer the question asked by user, Do not say unnecessary things.[QUESTION]"
+# System instructions for the models to follow
+system_instructions = "[SYSTEM] Your task is to Answer the question. Keep conversation very short, clear and concise. The expectation is that you will avoid introductions and start answering the query directly, Only answer the question asked by user, Do not say unnecessary things.[QUESTION]"
 
-def models(text, model="Mixtral 8x7B"): 
- 
+# Function to generate model responses
+def models(text, model="Mixtral 8x7B"):
  client = client_fn(model)
+ generate_kwargs = {
+ "max_new_tokens": 100,
+ "do_sample": True,
+ }
  
- generate_kwargs = dict(
- max_new_tokens=100,
- do_sample=True,
- )
- 
- formatted_prompt = system_instructions1 + text + "[ANSWER]"
+ formatted_prompt = f"{system_instructions} {text} [ANSWER]"
  stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
+ 
  output = ""
- for response in stream: 
- output+=response.token.text
- if output.endswith("<|assistant|>"):
- output = output[:-13]
- elif output.endswith("</s>"):
+ for response in stream:
+ output += response.token.text
+ if output.endswith("</s>"):
  output = output[:-4]
  return output
 
-description="""# Chat GO
+# Gradio interface description and configuration
+description = """# H GO
 ### Inspired from Google Go"""
 
-demo = gr.Interface(description=description,fn=models, inputs=["text", gr.Dropdown([ 'Mixtral 8x7B','Nous Hermes Mixtral 8x7B DPO','StarChat2 15b','Mistral 7B v0.3','Phi 3 mini', ], value="Mistral 7B v0.3", label="Select Model") ], outputs="text", live=True, batch=True, max_batch_size=10000)
+demo = gr.Interface(
+ description=description,
+ fn=models,
+ inputs=["text", gr.Dropdown(['Mixtral 8x7B', 'Nous Hermes Mixtral 8x7B DPO', 'StarChat2 15b', 'Mistral 7B v0.3', 'Phi 3 mini'], value="Mistral 7B v0.3", label="Select Model")],
+ outputs="text",
+ live=True,
+ batch=True,
+ max_batch_size=10000
+)
+
+# Queue and launch configuration for Gradio
 demo.queue(max_size=300000)
 demo.launch()