Switch to float16.

app.py

@@ -12,23 +12,23 @@ DEFAULT_MAX_NEW_TOKENS = 1024
 MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 DESCRIPTION = """\
-# Llama-2 13B Chat
+# Turkish LLaMA 8B Chat
 
-This Space demonstrates model [Llama-2-13b-chat](https://huggingface.co/meta-llama/Llama-2-13b-chat) by Meta, a Llama 2 model with 13B parameters fine-tuned for chat instructions. Feel free to play with it, or duplicate to run generations without a queue! If you want to run your own service, you can also [deploy the model on Inference Endpoints](https://huggingface.co/inference-endpoints).
+This Space demonstrates [Turkish-Llama-8b-DPO-v0.1](https://huggingface.co/ytu-ce-cosmos/Turkish-Llama-8b-DPO-v0.1) by YTU COSMOS Research Group, an 8B parameter model fine-tuned for Turkish language understanding and generation. Feel free to play with it, or duplicate to run generations without a queue!
 
-🔎 For more details about the Llama 2 family of models and how to use them with `transformers`, take a look [at our blog post](https://huggingface.co/blog/llama2).
+🔎 This model is the newest and most advanced iteration of CosmosLLama, developed by merging two distinctly trained CosmosLLaMa-Instruct DPO models.
 
-🔨 Looking for an even more powerful model? Check out the large [**70B** model demo](https://huggingface.co/spaces/ysharma/Explore_llamav2_with_TGI).
-🐇 For a smaller model that you can run on many GPUs, check our [7B model demo](https://huggingface.co/spaces/huggingface-projects/llama-2-7b-chat).
+🤖 The model is optimized for Turkish language tasks and can handle various text generation scenarios including conversations, instructions, and general text completion.
 
+💡 You can also try the model on the official demo page: [cosmos.yildiz.edu.tr/cosmosllama](https://cosmos.yildiz.edu.tr/cosmosllama)
 """
 
 LICENSE = """
 <p/>
 
 ---
-As a derivate work of [Llama-2-13b-chat](https://huggingface.co/meta-llama/Llama-2-13b-chat) by Meta,
-this demo is governed by the original [license](https://huggingface.co/spaces/huggingface-projects/llama-2-13b-chat/blob/main/LICENSE.txt) and [acceptable use policy](https://huggingface.co/spaces/huggingface-projects/llama-2-13b-chat/blob/main/USE_POLICY.md).
+This demo uses [Turkish-Llama-8b-DPO-v0.1](https://huggingface.co/ytu-ce-cosmos/Turkish-Llama-8b-DPO-v0.1) by YTU COSMOS Research Group,
+and is governed by the original llama3 license.
 """
 
 if not torch.cuda.is_available():
@@ -37,7 +37,11 @@ if not torch.cuda.is_available():
 
 if torch.cuda.is_available():
     model_id = "ytu-ce-cosmos/Turkish-Llama-8b-DPO-v0.1"
-    model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_4bit=True)
+    model = AutoModelForCausalLM.from_pretrained(
+        model_id,
+        device_map="auto",
+        torch_dtype=torch.float16,
+    )
     tokenizer = AutoTokenizer.from_pretrained(model_id)
     tokenizer.use_default_system_prompt = False
 
@@ -128,11 +132,11 @@ chat_interface = gr.ChatInterface(
     ],
     stop_btn=None,
     examples=[
-        ["Hello there! How are you doing?"],
-        ["Can you explain briefly to me what is the Python programming language?"],
-        ["Explain the plot of Cinderella in a sentence."],
-        ["How many hours does it take a man to eat a Helicopter?"],
-        ["Write a 100-word article on 'Benefits of Open-Source in AI research'"],
+        ["Merhaba! Nasılsın?"],
+        ["Python programlama dilini kısaca açıklayabilir misin?"],
+        ["Külkedisi masalının özetini bir cümlede anlat."],
+        ["Yapay zeka alanında açık kaynak kodun faydaları nelerdir?"],
+        ["İstanbul'un en ünlü turistik yerlerini sıralar mısın?"],
     ],
     cache_examples=False,
     type="messages",