Update app.py

app.py

@@ -1,64 +1,58 @@
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from datasets import load_dataset
+from transformers import Trainer, TrainingArguments
 import gradio as gr
-from huggingface_hub import InferenceClient
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+MODEL_NAME = "cahya/indogpt2-base"
 
+# Load tokenizer & model
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForCausalLM.from_pretrained(MODEL_NAME)
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+# Load dataset dari file lokal
+dataset = load_dataset("json", data_files={"train": "data.json"})
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+def tokenize_function(examples):
+    return tokenizer(examples["text"], truncation=True, max_length=512)
 
-    messages.append({"role": "user", "content": message})
+tokenized_dataset = dataset.map(tokenize_function, batched=True)
 
-    response = ""
+training_args = TrainingArguments(
+    output_dir="./indoGPT-finetuned",
+    per_device_train_batch_size=1,
+    gradient_accumulation_steps=8,
+    num_train_epochs=1,
+    logging_steps=10,
+    evaluation_strategy="no",
+    save_strategy="no",
+    learning_rate=5e-5,
+    fp16=True,
+)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+trainer = Trainer(
+    model=model,
+    args=training_args,
+    train_dataset=tokenized_dataset["train"],
+)
 
-        response += token
-        yield response
+def fine_tune():
+    trainer.train()
+    return "Fine-tuning selesai!"
 
+def chat(input_text):
+    inputs = tokenizer(input_text, return_tensors="pt")
+    outputs = model.generate(**inputs, max_length=100, do_sample=True, top_p=0.9, temperature=0.7)
+    return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
+with gr.Blocks() as demo:
+    gr.Markdown("# Chatbot IndoGPT Simple")
+    user_input = gr.Textbox(label="Ketik pesan kamu")
+    fine_tune_btn = gr.Button("Fine-tune model")
+    output = gr.Textbox(label="Balasan chatbot")
 
+    fine_tune_btn.click(fine_tune, outputs=output)
+    user_input.submit(chat, inputs=user_input, outputs=output)
 
 if __name__ == "__main__":
     demo.launch()