Upload 3 files

app.py

@@ -0,0 +1,115 @@
+import gradio as gr
+from huggingface_hub import InferenceClient
+import spaces #0.32.0
+import torch
+import os
+import platform
+import requests
+
+model = ""
+duration = None
+token = os.getenv('deepseekv2')
+provider = None #'fal-ai' #None #replicate # sambanova
+
+print(f"Is CUDA available: {torch.cuda.is_available()}")
+print(f"CUDA device: {torch.cuda.get_device_name(torch.cuda.current_device())}")
+print(f"CUDA version: {torch.version.cuda}")
+print(f"Python version: {platform.python_version()}")
+print(f"Pytorch version: {torch.__version__}")
+print(f"Gradio version: {gr. __version__}")
+# print(f"HFhub version: {huggingface_hub.__version__}")
+
+
+"""
+Packages ::::::::::
+Is CUDA available: True
+CUDA device: NVIDIA A100-SXM4-80GB MIG 3g.40gb
+CUDA version: 12.1
+Python version: 3.10.13
+Pytorch version: 2.4.0+cu121
+Gradio version: 5.0.1
+"""
+
+
+def choose_model(model_name):
+    if model_name == "DeepSeek-R1-Distill-Qwen-1.5B":
+        model = "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B"
+
+    elif model_name == "DeepSeek-R1-Distill-Qwen-32B":
+        model = "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"
+        
+    elif model_name == "Llama3-8b-Instruct":    
+        model = "meta-llama/Meta-Llama-3-8B-Instruct"
+
+    elif model_name == "Llama3.1-8b-Instruct":
+        model = "meta-llama/Llama-3.1-8B-Instruct"
+
+    elif model_name == "Llama2-13b-chat":
+        model = "meta-llama/Llama-2-13b-chat-hf"
+
+    elif model_name == "Gemma-2-2b":
+        model = "google/gemma-2-2b-it"
+
+    elif model_name == "Gemma-7b":
+        model = "google/gemma-7b"
+    
+    elif model_name == "Mixtral-8x7B-Instruct":
+        model = "mistralai/Mixtral-8x7B-Instruct-v0.1"
+
+    elif model_name == "Microsoft-phi-2":
+        model = "microsoft/phi-2"
+
+    elif model_name == "Qwen2.5-Coder-32B-Instruct":
+        model = "Qwen/Qwen2.5-Coder-32B-Instruct"
+
+    else:    # default to zephyr if no model chosen
+        model = "HuggingFaceH4/zephyr-7b-beta"
+    
+    return model
+    
+
+@spaces.GPU(duration=duration)
+def respond(message, history: list[tuple[str, str]], model, system_message, max_tokens, temperature, top_p):
+
+    print(model)
+    model_name = choose_model(model)
+
+    client = InferenceClient(model_name, provider=provider, token=os.getenv('deepseekv2'))
+    
+    messages = [{"role": "system", "content": system_message}]
+
+    for val in history:
+        if val[0]:
+            messages.append({"role": "user", "content": val[0]})
+        if val[1]:
+            messages.append({"role": "assistant", "content": val[1]})
+
+    messages.append({"role": "user", "content": message})
+
+    response = ""
+
+    for message in client.chat_completion(messages, max_tokens=max_tokens, stream=True, temperature=temperature, top_p=top_p):
+        token = message.choices[0].delta.content
+
+        response += token
+        yield response
+
+    
+demo = gr.ChatInterface(
+    respond,
+    title="Ask me anything",
+    description="Hi there! I am your friendly AI chatbot. Choose from different language models under the Additional Inputs tab below.",
+    examples=[["Explain quantum computing"], ["Explain forex trading"], ["What is the capital of China?"], ["Make a poem about nature"]],
+    additional_inputs=[
+        gr.Dropdown(["DeepSeek-R1-Distill-Qwen-1.5B", "DeepSeek-R1-Distill-Qwen-32B", "Gemma-2-2b", "Gemma-7b", "Llama2-13b-chat", "Llama3-8b-Instruct", "Llama3.1-8b-Instruct", "Microsoft-phi-2", "Mixtral-8x7B-Instruct", "Qwen2.5-Coder-32B-Instruct", "Zephyr-7b-beta"], label="Select Model"),
+        gr.Textbox(value="You are a friendly and helpful Chatbot, be concise and straight to the point, avoid excessive reasoning.", label="System message"),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)")
+        
+    ]
+)
+
+
+if __name__ == "__main__":
+    demo.launch(share=True)

providers

@@ -0,0 +1,3 @@
+provider (str, optional) — Name of the provider to use for inference. Can be "black-forest-labs", "fal-ai", "fireworks-ai", "hf-inference",
+"hyperbolic", "nebius", "novita", "replicate", “sambanova”or“together”. defaults to hf-inference (Hugging Face Serverless Inference API).
+If model is a URL or base_urlis passed, thenprovider` is not used.

requirements.txt

@@ -0,0 +1,5 @@
+huggingface_hub==0.28.1
+--extra-index-url https://download.pytorch.org/whl/cu124
+torch==2.4.0
+spaces
+gradio==5.12.0