update

app.py

@@ -194,6 +194,7 @@ def evaluate(
         )
     input_ids = inputs["input_ids"].to(device)
     generation_config = GenerationConfig(
+        do_sample=True,
         temperature=temperature,
         top_p=top_p,
         top_k=top_k,
@@ -258,7 +259,6 @@ description = (
     "It is trained on the [izumi-lab/llm-japanese-dataset](https://huggingface.co/datasets/izumi-lab/llm-japanese-dataset) dataset.  "
     "For more information, please visit [the project's website](https://llm.msuzuki.me).  "
     "This model can output up to 256 tokens, but the maximum number of tokens is 227 due to the GPU memory limit of HuggingFace Space.  "
-    "Currently temperature is ineffective because num_beam is set to 1 due to lack of GPU memory.  "
     "It takes about **1 minute** to output. When access is concentrated, the operation may become slow."
 )
 with gr.Blocks(
@@ -289,7 +289,7 @@ with gr.Blocks(
                 value=0.7,
                 step=0.05,
                 interactive=True,
-                label="Temperature (Currently ineffective due to insufficient GPU memory)",
+                label="Temperature",
             )
             max_tokens = gr.Slider(
                 minimum=20,