Add Seq2Seq example using T5

examples/seq2seq_quickstart.py

@@ -0,0 +1,37 @@
+from simpletransformers.seq2seq import Seq2SeqModel, Seq2SeqArgs
+import pandas as pd
+
+# Sample data: English to French
+train_data = [
+    ["translate English to French: Hello, how are you?", "Bonjour, comment ça va ?"],
+    ["translate English to French: What is your name?", "Comment tu t'appelles ?"],
+]
+
+train_df = pd.DataFrame(train_data, columns=["input_text", "target_text"])
+
+# Define model
+model_args = Seq2SeqArgs()
+model_args.num_train_epochs = 2
+model_args.train_batch_size = 2
+model_args.max_seq_length = 64
+model_args.eval_batch_size = 2
+model_args.save_eval_checkpoints = False
+model_args.save_model_every_epoch = False
+model_args.use_multiprocessing = False
+model_args.no_cache = True
+model_args.overwrite_output_dir = True
+
+model = Seq2SeqModel(
+    encoder_decoder_type="t5",
+    encoder_decoder_name="t5-small",
+    args=model_args,
+    use_cuda=False  # Change to True if you have a GPU
+)
+
+# Train the model
+model.train_model(train_df)
+
+# Predict
+inputs = ["translate English to French: I am learning AI."]
+predictions = model.predict(inputs)
+print(predictions)