Update README.md

README.md

@@ -17,6 +17,54 @@ language:
 - **License:** apache-2.0
 - **Finetuned from model :** unsloth/Qwen3-4B-Base
 
-This qwen3 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
+```
+!pip install transformers accelerate peft
+```
 
-[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
+
+usage
+```python
+import torch
+print(torch.cuda.is_available())
+
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+# Model ve tokenizer yükle
+model_name = "Sengil/qwen3-4b-turkish-summarizer"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    device_map="auto",         # accelerate devreye girer
+    torch_dtype=torch.float16  # optimize belleği
+)
+model.eval()
+
+# Mesaj formatı
+messages = [
+    {"role": "system", "content": "Sen bir özetleyicisin. Sana verilen metni özetle."},
+    {"role": "user", "content": "text. . ."},
+]
+
+# Chat template ile prompt oluştur
+prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+
+# Tokenizer ile input tensor'ları oluştur
+inputs = tokenizer(prompt, return_tensors="pt")  # Cihaza atama yapma!
+
+# İnferans (generate) işlemi
+with torch.no_grad():
+    outputs = model.generate(
+        input_ids=inputs["input_ids"].to(model.device),
+        attention_mask=inputs["attention_mask"].to(model.device),
+        max_new_tokens=128,
+        do_sample=False,
+        temperature=0.7,
+    )
+
+# Çıktıyı çöz
+summary = tokenizer.decode(outputs[0], skip_special_tokens=True)
+print("\n📝 Özet:", summary)
+
+```