Update Llama 3.1 8B robot planning space with improvements

README.md

@@ -13,18 +13,92 @@ hardware: t4-medium
 
 # 🤖 Robot Task Planning - Llama 3.1 8B
 
-Fine-tuned Llama 3.1 8B model for robot task planning using QLoRA technique.
+This Space demonstrates a fine-tuned version of Meta's **Llama 3.1 8B** model specialized for **robot task planning** using QLoRA (4-bit quantization + LoRA) technique.
 
-## Model
-[YongdongWang/llama-3.1-8b-dart-qlora](https://huggingface.co/YongdongWang/llama-3.1-8b-dart-qlora)
+## 🎯 Purpose
 
-## Features
-- Natural language to robot task conversion
-- Multi-robot coordination
-- Real-time task generation
-- Optimized with 4-bit quantization
+Convert natural language commands into structured task sequences for construction robots including:
+- **Excavators** - Digging, loading, positioning
+- **Dump Trucks** - Material transport, loading, unloading  
+- **Multi-robot Coordination** - Complex task dependencies
 
-## Usage
-Input robot commands and get structured task sequences for excavators, dump trucks, and other construction robots.
+## 🔗 Model
 
-Loading time: ~3-5 minutes on first startup.
+**Fine-tuned Model**: [YongdongWang/llama-3.1-8b-dart-qlora](https://huggingface.co/YongdongWang/llama-3.1-8b-dart-qlora)
+
+**Base Model**: [meta-llama/Llama-3.1-8B](https://huggingface.co/meta-llama/Llama-3.1-8B)
+
+## ✨ Features
+
+- 🎮 **Interactive Chat Interface** - Real-time robot command processing
+- ⚙️ **Configurable Generation** - Adjust temperature, top-p, max tokens
+- 📝 **Example Commands** - Pre-built scenarios to get started
+- 🚀 **Optimized Performance** - 4-bit quantization for efficient inference
+- 📊 **Structured Output** - JSON-formatted task sequences
+
+## 🚀 Usage
+
+1. **Input**: Natural language robot commands
+   ```
+   "Deploy Excavator 1 to Soil Area 1 for excavation"
+   ```
+
+2. **Output**: Structured task sequences
+   ```json
+   {
+     "tasks": [
+       {
+         "robot": "Excavator_1",
+         "action": "move_to",
+         "target": "Soil_Area_1",
+         "duration": 30
+       },
+       {
+         "robot": "Excavator_1", 
+         "action": "excavate",
+         "target": "Soil_Area_1",
+         "duration": 120
+       }
+     ]
+   }
+   ```
+
+## 🛠️ Technical Details
+
+- **Architecture**: Llama 3.1 8B + QLoRA adapters
+- **Quantization**: 4-bit (NF4) with double quantization
+- **Framework**: Transformers + PEFT + BitsAndBytesConfig
+- **Interface**: Gradio 4.44.0
+- **Hardware**: T4 Medium (16GB VRAM)
+
+## ⚡ Performance Notes
+
+- **First Load**: 3-5 minutes (model downloading + loading)
+- **Subsequent Generations**: ~2-10 seconds per response
+- **Memory Usage**: ~8GB VRAM with 4-bit quantization
+- **Context Length**: Up to 2048 tokens
+
+## 📚 Example Commands
+
+Try these robot commands:
+
+- `"Deploy Excavator 1 to Soil Area 1 for excavation"`
+- `"Send Dump Truck 1 to collect material, then unload at storage"`
+- `"Coordinate multiple excavators across different areas"`
+- `"Create evacuation sequence for all robots from dangerous zone"`
+
+## 🔬 Research Applications
+
+This model demonstrates:
+- **Natural Language → Robot Planning** translation
+- **Multi-agent Task Coordination** 
+- **Efficient LLM Fine-tuning** with QLoRA
+- **Real-time Robot Command Processing**
+
+## 📄 License
+
+This project uses Meta's Llama 3.1 license. Please review the license terms before use.
+
+## 🤝 Contributing
+
+For issues, improvements, or questions about the model, please visit the [model repository](https://huggingface.co/YongdongWang/llama-3.1-8b-dart-qlora).

app.py

@@ -57,13 +57,13 @@ def load_model():
         print(f"❌ Model loading failed: {load_error}")
         return None, None
 
-# 全局变量
+# 全局变量存储模型
 model = None
 tokenizer = None
 model_loading = False
 
 def initialize_model():
-    """初始化模型"""
+    """初始化模型 - 延迟加载"""
     global model, tokenizer, model_loading
     
     if model is not None and tokenizer is not None:
@@ -85,7 +85,7 @@ def generate_response(prompt, max_tokens=200, temperature=0.7, top_p=0.9):
         if model_loading:
             return "🔄 Model is loading, please wait a few minutes and try again..."
         else:
-            return "❌ Model failed to load. Please check the Space logs."
+            return "❌ Model failed to load. Please check the Space logs or try restarting."
     
     try:
         # 格式化输入
@@ -123,7 +123,7 @@ def generate_response(prompt, max_tokens=200, temperature=0.7, top_p=0.9):
         elif len(response) > len(formatted_prompt):
             response = response[len(formatted_prompt):].strip()
         
-        return response if response else "❌ No response generated. Please try again."
+        return response if response else "❌ No response generated. Please try again with a different prompt."
     
     except Exception as generation_error:
         return f"❌ Generation Error: {str(generation_error)}"
@@ -156,55 +156,112 @@ with gr.Blocks(
     gr.Markdown("""
     # 🤖 Llama 3.1 8B - Robot Task Planning
     
-    Fine-tuned version of Meta's Llama 3.1 8B for **robot task planning** using QLoRA.
+    This is a fine-tuned version of Meta's Llama 3.1 8B model specialized for **robot task planning** using QLoRA technique.
+    
+    **Capabilities**: Convert natural language robot commands into structured task sequences for excavators, dump trucks, and other construction robots.
     
     **Model**: [YongdongWang/llama-3.1-8b-dart-qlora](https://huggingface.co/YongdongWang/llama-3.1-8b-dart-qlora)
     
-    ⚠️ **First load takes 3-5 minutes**
+    ⚠️ **Note**: Model loading may take 3-5 minutes on first startup. Please be patient.
     """)
     
     with gr.Row():
         with gr.Column(scale=3):
             chatbot = gr.Chatbot(
-                label="🤖 Task Planning Results",
+                label="Task Planning Results",
                 height=500,
+                show_label=True,
+                container=True,
+                bubble_full_width=False,
                 show_copy_button=True
             )
             
             msg = gr.Textbox(
                 label="Robot Command",
-                placeholder="e.g., 'Deploy Excavator 1 to Soil Area 1'...",
-                lines=2
+                placeholder="Enter robot task command (e.g., 'Deploy Excavator 1 to Soil Area 1')...",
+                lines=2,
+                max_lines=5,
+                show_label=True,
+                container=True
             )
             
             with gr.Row():
-                send_btn = gr.Button("🚀 Generate", variant="primary")
-                clear_btn = gr.Button("🗑️ Clear", variant="secondary")
+                send_btn = gr.Button("🚀 Generate Tasks", variant="primary", size="sm")
+                clear_btn = gr.Button("🗑️ Clear", variant="secondary", size="sm")
         
         with gr.Column(scale=1):
-            gr.Markdown("### ⚙️ Settings")
+            gr.Markdown("### ⚙️ Generation Settings")
+            
+            max_tokens = gr.Slider(
+                minimum=50,
+                maximum=500,
+                value=200,
+                step=10,
+                label="Max Tokens",
+                info="Maximum number of tokens to generate"
+            )
+            
+            temperature = gr.Slider(
+                minimum=0.1,
+                maximum=2.0,
+                value=0.7,
+                step=0.1,
+                label="Temperature",
+                info="Controls randomness (lower = more focused)"
+            )
+            
+            top_p = gr.Slider(
+                minimum=0.1,
+                maximum=1.0,
+                value=0.9,
+                step=0.05,
+                label="Top-p",
+                info="Nucleus sampling threshold"
+            )
             
-            max_tokens = gr.Slider(50, 500, 200, label="Max Tokens")
-            temperature = gr.Slider(0.1, 2.0, 0.7, step=0.1, label="Temperature")
-            top_p = gr.Slider(0.1, 1.0, 0.9, step=0.05, label="Top-p")
+            gr.Markdown("""
+            ### 📊 Model Status
+            The model will load automatically on first use.
+            Loading time: ~3-5 minutes
+            """)
     
-    # 示例
+    # 示例对话
     gr.Examples(
         examples=[
             ["Deploy Excavator 1 to Soil Area 1 for excavation."],
-            ["Send Dump Truck 1 to collect material and unload at storage."],
-            ["Move all robots to avoid dangerous Puddle 1."],
-            ["Coordinate multiple excavators across different areas."],
-            ["Create evacuation sequence for all robots."],
+            ["Send Dump Truck 1 to collect material from Excavator 1, then unload at storage area."],
+            ["Move all robots to avoid Puddle 1 after inspection."],
+            ["Deploy multiple excavators to different soil areas simultaneously."],
+            ["Coordinate dump trucks to transport materials from excavation site to storage."],
+            ["Send robot to inspect rock area, then avoid with all other robots if dangerous."],
+            ["Return all robots to start position after completing tasks."],
+            ["Create a sequence: excavate, load, transport, unload, repeat."]
         ],
         inputs=msg,
-        label="💡 Try these examples"
+        label="💡 Example Robot Commands"
     )
     
     # 事件处理
-    msg.submit(chat_interface, [msg, chatbot, max_tokens, temperature, top_p], [chatbot, msg])
-    send_btn.click(chat_interface, [msg, chatbot, max_tokens, temperature, top_p], [chatbot, msg])
-    clear_btn.click(lambda: ([], ""), outputs=[chatbot, msg])
+    msg.submit(
+        chat_interface,
+        inputs=[msg, chatbot, max_tokens, temperature, top_p],
+        outputs=[chatbot, msg]
+    )
+    
+    send_btn.click(
+        chat_interface,
+        inputs=[msg, chatbot, max_tokens, temperature, top_p],
+        outputs=[chatbot, msg]
+    )
+    
+    clear_btn.click(
+        lambda: ([], ""),
+        outputs=[chatbot, msg]
+    )
 
 if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", server_port=7860)
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        show_error=True
+    )