Upload 3 files

README.md

@@ -1,12 +1,26 @@
----
-title: Ai2
-emoji: 📚
-colorFrom: green
-colorTo: yellow
-sdk: gradio
-sdk_version: 6.9.0
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+---
+title: Qwen3.5 Reasoning Chat (Claude Distilled)
+emoji: 🧠
+colorFrom: blue
+colorTo: indigo
+sdk: gradio
+app_file: app.py
+pinned: false
+license: apache-2.0
+---
+
+# Qwen3.5 Reasoning Chat (Claude 4.6 Distilled)
+
+This project runs the **Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2-GGUF** model using `llama-cpp-python`.
+
+It is specifically tuned to follow a **Chain-of-Thought (CoT)** reasoning process using `<think>` tags, distilled from Claude 4.6 Opus trajectories.
+
+## Features
+- **GGUF Optimized:** Runs efficiently on CPU.
+- **Reasoning First:** Designed to "think" before answering.
+- **No API Key Required:** Runs the model directly within the Space environment.
+
+## Deployment
+1. Log in: `huggingface-cli login`
+2. Create Space: `huggingface-cli repo create qwen3.5-reasoning --repo-type space --sdk gradio`
+3. Upload: `huggingface-cli upload <username>/qwen3.5-reasoning . --repo-type space`

app.py

@@ -0,0 +1,50 @@
+import gradio as gr
+from huggingface_hub import hf_hub_download
+from llama_cpp import Llama
+import os
+
+# 1. Configuration - Specify the model and the specific GGUF file
+model_id = "Jackrong/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2-GGUF"
+filename = "Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2-Q4_K_M.gguf"
+
+print(f"Downloading model: {filename}...")
+model_path = hf_hub_download(repo_id=model_id, filename=filename)
+
+# 2. Initialize the model
+# n_ctx is the context window. 2048 is a safe starting point for free Hugging Face Spaces.
+llm = Llama(model_path=model_path, n_ctx=2048, n_threads=os.cpu_count())
+
+def generate_response(message, history):
+    # Construct the prompt for reasoning models
+    # They typically look for a specific structure to trigger <think> tags.
+    prompt = f"<|im_start|>system\nYou are a helpful assistant with advanced reasoning capabilities. Use <think> tags for your internal logic.<|im_end|>\n"
+    
+    for user_msg, assistant_msg in history:
+        prompt += f"<|im_start|>user\n{user_msg}<|im_end|>\n"
+        prompt += f"<|im_start|>assistant\n{assistant_msg}<|im_end|>\n"
+    
+    prompt += f"<|im_start|>user\n{message}<|im_end|>\n<|im_start|>assistant\n<think>\n"
+    
+    output = llm(
+        prompt,
+        max_tokens=1024,
+        stop=["<|im_end|>"],
+        stream=True
+    )
+    
+    response = ""
+    for chunk in output:
+        delta = chunk['choices'][0]['text']
+        response += delta
+        yield response
+
+# 3. Create the UI
+demo = gr.ChatInterface(
+    generate_response,
+    title="Qwen 3.5 Reasoning Chat (Claude Distilled)",
+    description="This Space runs the Qwen3.5-9B reasoning model distilled from Claude 4.6 Opus logic.",
+    examples=["Solve the riddle: What has keys but can't open locks?", "Explain quantum entanglement in simple terms."],
+)
+
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,3 @@
+gradio
+huggingface_hub
+llama-cpp-python