Update app.py

app.py

@@ -1,27 +1,37 @@
-import gradio as gr
+
 import os
 os.system("pip install torch sentencepiece transformers Xformers accelerate")
+import gradio as gr
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
+# モデルとトークナイザの初期化
 model = AutoModelForCausalLM.from_pretrained("cyberagent/open-calm-1b", device_map="auto", torch_dtype=torch.float16)
 tokenizer = AutoTokenizer.from_pretrained("cyberagent/open-calm-1b")
 
-def generate_text(prompt, max_new_tokens, do_sample, temperature, top_p, repetition_penalty, pad_token_id):
-  inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
-  with torch.no_grad():
-    tokens = model.generate(
-        **inputs,
-        max_new_tokens=max_new_tokens,
-        do_sample=do_sample,
-        temperature=temperature,
-        top_p=top_p,
-        repetition_penalty=repetition_penalty,
-        pad_token_id=pad_token_id,
-    )
-    
-  output = tokenizer.decode(tokens[0], skip_special_tokens=True)
-  return output
+# 推論用の関数
+def generate_text(input_text):
+    inputs = tokenizer(input_text, return_tensors="pt").to(model.device)
+    with torch.no_grad():
+        tokens = model.generate(
+            **inputs,
+            max_new_tokens=64,
+            do_sample=True,
+            temperature=0.7,
+            top_p=0.9,
+            repetition_penalty=1.05,
+            pad_token_id=tokenizer.pad_token_id,
+        )
+    output = tokenizer.decode(tokens[0], skip_special_tokens=True)
+    return output
+
+# 入力と出力のインターフェースを作成
+input_text = gr.inputs.Textbox(lines=2, label="入力テキスト")
+output_text = gr.outputs.Textbox(label="生成されたテキスト")
+
+# インターフェースを作成
+iface = gr.Interface(fn=generate_text, inputs=input_text, outputs=output_text)
+
+# GUIを起動
+iface.launch()
 
-app = gr.Interface(generate_text, inputs=[gr.inputs(label="Prompt", type="text"), gr.IntSlider(label="Max new tokens", min=1, max=1024, step=1), gr.Checkbox(label="Do sample"), gr.FloatSlider(label="Temperature", min=0.1, max=1.0, step=0.1), gr.FloatSlider(label="Top P", min=0.0, max=1.0, step=0.01), gr.FloatSlider(label="Repetition penalty", min=0.0, max=2.0, step=0.1), gr.IntSlider(label="Pad token ID", min=0, max=1023, step=1)], outputs=[gr.Output(label="Output", type="text")])
-app.launch()