Create train_model.py

train_model.py

@@ -0,0 +1,29 @@
+from transformers import LayoutLMForTokenClassification, Trainer, TrainingArguments
+from datasets import load_dataset
+
+# Wczytanie przygotowanego zbioru danych
+dataset = load_dataset("json", data_files="training_data.json")
+
+# Ładowanie modelu LayoutLM do dostrajania
+model = LayoutLMForTokenClassification.from_pretrained("microsoft/layoutlmv3-base", num_labels=5)
+
+training_args = TrainingArguments(
+    output_dir="./layoutlmv3_finetuned",
+    per_device_train_batch_size=4,
+    per_device_eval_batch_size=4,
+    num_train_epochs=5,
+    evaluation_strategy="epoch",
+    save_strategy="epoch"
+)
+
+trainer = Trainer(
+    model=model,
+    args=training_args,
+    train_dataset=dataset["train"],
+    eval_dataset=dataset["test"]
+)
+
+trainer.train()
+
+# Zapisanie modelu
+model.save_pretrained("./layoutlmv3_finetuned")