Update model to unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF

- Changed current_model in backend_service.py
- Updated ChatCompletionRequest default model
- Updated all test files to use new model
- Updated requirements.txt comment
- DeepSeek R1-0528 is a more modern and capable model than the previous gemma model

backend_service.py

@@ -75,7 +75,7 @@ class ChatMessage(BaseModel):
         return v
 
 class ChatCompletionRequest(BaseModel):
-    model: str = Field(default="gemma-3n-E4B-it-GGUF", description="The model to use for completion")
+    model: str = Field(default="unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF", description="The model to use for completion")
     messages: List[ChatMessage] = Field(..., description="List of messages in the conversation")
     max_tokens: Optional[int] = Field(default=512, ge=1, le=2048, description="Maximum tokens to generate")
     temperature: Optional[float] = Field(default=0.7, ge=0.0, le=2.0, description="Sampling temperature")
@@ -124,7 +124,7 @@ class CompletionRequest(BaseModel):
 # Global variables for model management
 inference_client: Optional[InferenceClient] = None
 image_text_pipeline = None  # type: ignore
-current_model = "gemma-3n-E4B-it-GGUF"
+current_model = "unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF"
 vision_model = "Salesforce/blip-image-captioning-base"  # Working model for image captioning
 tokenizer = None
 

requirements.txt

@@ -5,7 +5,7 @@ torch>=2.0.0
 Pillow>=10.0.0
 accelerate>=0.24.0
 requests>=2.31.0
-# NOTE: GGUF models like 'gemma-3n-E4B-it-GGUF' must be downloaded manually or referenced from HuggingFace, not pip-installed.
+# NOTE: GGUF models like 'unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF' must be downloaded manually or referenced from HuggingFace, not pip-installed.
 fastapi>=0.100.0
 uvicorn[standard]>=0.23.0
 pydantic>=2.0.0

test_hf_api.py

@@ -5,7 +5,7 @@ API_URL = "https://cong182-firstai.hf.space/v1/chat/completions"
 
 # Example payload for OpenAI-compatible chat completion
 payload = {
-    "model": "gemma-3n-E4B-it-GGUF",
+    "model": "unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF",
     "messages": [
         {"role": "system", "content": "You are a helpful assistant."},
         {"role": "user", "content": "Hello, who won the world cup in 2018?"}

test_multimodal.py

@@ -42,7 +42,7 @@ def test_multimodal():
     print("🖼️ Testing multimodal chat completion...")
     
     payload = {
-        "model": "unsloth/gemma-3n-E4B-it-GGUF",
+        "model": "unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF",
         "messages": [
             {
                 "role": "user",

test_pipeline.py

@@ -19,7 +19,7 @@ def test_pipeline_availability():
         models_to_try = [
             "Salesforce/blip-image-captioning-base",  # More common model
             "microsoft/git-base-textcaps",  # Alternative model
-            "unsloth/gemma-3n-E4B-it-GGUF"  # Original model
+            "unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF"  # Updated model
         ]
         
         for model_name in models_to_try: