DOCS: update README for GraLoRA finetuning with correct SFTTrainer integration

examples/gralora_finetuning/README.md

@@ -13,28 +13,34 @@ With respect to your standard PEFT training procedure with LoRA, simply swap you
 
 ```python
 import torch
-from peft import GraloraConfig, get_peft_model
-from transformers import AutoTokenizer, AutoModelForCausalLM, Trainer
+from peft import GraloraConfig
+from transformers import AutoTokenizer, AutoModelForCausalLM
 from datasets import load_dataset
+from trl import SFTTrainer, SFTConfig
 
 model = AutoModelForCausalLM.from_pretrained("huggyllama/llama-7b", device_map="auto")
 tokenizer = AutoTokenizer.from_pretrained("huggyllama/llama-7b")
 dataset = load_dataset("timdettmers/openassistant-guanaco", split="train")
 gralora_config = GraloraConfig()
-peft_model = get_peft_model(model, gralora_config)
-trainer = transformers.Trainer(
-    model=peft_model,
+
+trainer = SFTTrainer(
+    model=model,
     train_dataset=dataset,
-    dataset_text_field="text",
-    max_seq_length=2048,
-    tokenizer=tokenizer,
+    processing_class=tokenizer,
+    peft_config=gralora_config,
+    args=SFTConfig(
+        output_dir="./gralora-llama-7b",
+        max_length=2048,
+        dataset_text_field="text",
+        per_device_train_batch_size=2,
+    ),
 )
 trainer.train()
-peft_model.save_pretrained("gralora-llama-3-8b")
+trainer.model.save_pretrained("gralora-llama-7b")
 ```
 
 Run the finetuning script simply by running:
-```python
+```sh
 python examples/gralora_finetuning/gralora_finetuning.py --base_model meta-llama/Meta-Llama-3-8B --data_path timdettmers/openassistant-guanaco
 ```
 
@@ -51,12 +57,9 @@ model = AutoModelForCausalLM.from_pretrained(
 peft_model = PeftModel.from_pretrained(model, "gralora-llama-3-8b")
 ```
 
-## Additonal Notes
+## Additional Notes
 While `gralora_k` is set to 2 for default, you can increase this value to create more fine-grained adapters. `gralora_k` of 4 is recommended when the total rank (`r + hybrid_r`) is 64 or higher.
 
-
-
-
 ## Citation
 ```
 @misc{jung2025graloragranularlowrankadaptation,