Update app.py

app.py

@@ -10,8 +10,8 @@ import spaces
 
 # Load model directly
 device = "cuda" if torch.cuda.is_available() else "cpu"
-tokenizer = AutoTokenizer.from_pretrained("Navid-AI/Mulhem-1-Mini", token=os.getenv("HF_TOKEN"))
-model = AutoModelForCausalLM.from_pretrained("Navid-AI/Mulhem-1-Mini", torch_dtype=torch.bfloat16, attn_implementation="flash_attention_2", token=os.getenv("HF_TOKEN")).to(device)
+tokenizer = AutoTokenizer.from_pretrained("Navid-AI/Yehia-R1-3B", token=os.getenv("HF_TOKEN"))
+model = AutoModelForCausalLM.from_pretrained("Navid-AI/Yehia-R1-3B", torch_dtype=torch.bfloat16, attn_implementation="flash_attention_2", token=os.getenv("HF_TOKEN")).to(device)
 streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
 
 @spaces.GPU
@@ -23,7 +23,6 @@ def respond(
     max_tokens,
     temperature,
     repetition_penalty,
-    top_p,
 ):
     messages = [{"role": "system", "content": system_message}]
 
@@ -36,7 +35,7 @@ def respond(
     messages.append({"role": "user", "content": message})
     print(messages)
     inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True, enable_reasoning=enable_reasoning, return_dict=True).to(device)
-    generation_kwargs = dict(inputs, streamer=streamer, max_new_tokens=max_tokens, temperature=temperature, top_p=top_p, repetition_penalty=repetition_penalty)
+    generation_kwargs = dict(inputs, streamer=streamer, max_new_tokens=max_tokens, temperature=temperature, repetition_penalty=repetition_penalty, eos_token_id=tokenizer.convert_tokens_to_ids("<|im_end|>"))
     thread = Thread(target=model.generate, kwargs=generation_kwargs)
 
     thread.start()
@@ -51,16 +50,9 @@ demo = gr.ChatInterface(
     additional_inputs=[
         gr.Checkbox(label="Enable reasoning", value=False),
         gr.Textbox(value="أنت مُلهم. ذكاء اصطناعي تم إنشاؤه من شركة نفيد لإلهام وتحفيز المستخدمين على التعلّم، النمو، وتحقيق أهدافهم.", label="System message"),
-        gr.Slider(minimum=1, maximum=8192, value=2048, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.1, step=0.1, label="Temperature"),
-        gr.Slider(minimum=0.1, maximum=2.0, value=1.25, step=0.05, label="Repetition penalty"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
+        gr.Slider(minimum=1, maximum=8192, value=4096, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.6, step=0.1, label="Temperature"),
+        gr.Slider(minimum=0.1, maximum=2.0, value=1.1, step=0.05, label="Repetition penalty"),
     ],
 )