Added specific use for my model

app.py

@@ -1,63 +1,58 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from peft import PeftModel
+import torch
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
-
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    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]})
+# Load the base model and LoRA adapter
+model_name = "unsloth/llama-3-8b-bnb-4bit"  # Replace with your base model
+adapter_model_name = "your-hf-username/your-lora-model"  # Replace with your LoRA adapter
 
-    messages.append({"role": "user", "content": message})
+# Load the base model and tokenizer
+base_model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16, device_map="auto")
+model = PeftModel.from_pretrained(base_model, adapter_model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
 
-    response = ""
+# Define the Alpaca-style prompt
+alpaca_prompt = """### Instruction:
+{instruction}
 
-    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
+### Input:
+{input}
 
-        response += token
-        yield response
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
+### Output:
+{output}
 """
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", 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)",
-        ),
-    ],
-)
 
+# Define the function to generate text
+def generate_response(instruction, input_text):
+    # Format the prompt with the instruction and input text
+    prompt = alpaca_prompt.format(
+        instruction=instruction,
+        input=input_text,
+        output=""  # Leave output blank for generation
+    )
+    
+    # Tokenize the prompt and move it to the GPU
+    inputs = tokenizer([prompt], return_tensors="pt").to("cuda")
+    
+    # Generate the response from the model
+    outputs = model.generate(**inputs, max_new_tokens=64, use_cache=True)
+    
+    # Decode the output into human-readable text
+    generated_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
+    
+    return generated_text
+
+# Gradio Interface with two inputs: Instruction and Input Text
+iface = gr.Interface(
+    fn=generate_response, 
+    inputs=[
+        gr.inputs.Textbox(lines=2, placeholder="Enter the instruction here..."), 
+        gr.inputs.Textbox(lines=5, placeholder="Enter the input text here...")
+    ], 
+    outputs="text", 
+    title="Alpaca-Style Instruction-Input-Output Model"
+)
 
-if __name__ == "__main__":
-    demo.launch()
+# Launch the Gradio app
+iface.launch()

requirements.txt

@@ -1 +1,5 @@
-huggingface_hub==0.22.2
+huggingface_hub==0.22.2
+transformers
+torch
+gradio
+peft