gguf to transformers model

- Issue with GGUF, changed to transformers

app.py

@@ -16,23 +16,18 @@ class RequestData(BaseModel):
 
 
 def load_model():
-    chat_handler = MoondreamChatHandler.from_pretrained(
-      repo_id="vikhyatk/moondream2",
-      filename="*mmproj*",
+    model_id = "models"
+    revision = "2024-08-26"
+    model = AutoModelForCausalLM.from_pretrained(
+        model_id, trust_remote_code=True, revision=revision
     )
-    
-    llm = Llama.from_pretrained(
-      repo_id="vikhyatk/moondream2",
-      filename="*text-model*",
-      chat_handler=chat_handler,
-      n_ctx=2048, # n_ctx should be increased to accommodate the image embedding
-    )
-    return llm
+    tokenizer = AutoTokenizer.from_pretrained(model_id, revision=revision)
+    return model, tokenizer
 
 
 
 llm = load_model()
-
+print("INFO: Model & Tokenizer loaded!")
 
 @app.get("/")
 def greet_json():
@@ -43,27 +38,23 @@ def greet_json():
 def query(data: RequestData):
     prompt = data.prompt
     image = data.image
-
+    print(f"INFO: prompt - {prompt}")
+    
     try:
         # decode base64 to image
-        data_uri = f"""data:image/png;base64,{image}"""
-        messages = [
-            {"role": "system", "content": "You are an assistant who perfectly describes images."},
-            {
-                "role": "user",
-                "content": [
-                    {"type": "image_url", "image_url": {"url": data_uri }},
-                    {"type" : "text", "text": f"""{prompt}"""}
-                ]
-            }
-        ]
-
-        response = llm.create_chat_completion(
-            messages = messages
-        )
-        output = response['choices'][0]['message']['content']
-
-        return {"response": str(output)}
+        image = base64.b64decode(image)
+
+        with NamedTemporaryFile(delete=True, suffix=".png") as temp_image:
+            temp_image.write(image)
+            temp_image.flush()
+
+            image = Image.open(temp_image.name)
+
+            image = Image.open("<IMAGE_PATH>")
+            enc_image = MODEL.encode_image(image)
+            response = MODEL.answer_question(enc_image, str(prompt), TOKENIZER)
+
+            return {"response": str(response)}
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))