Fix Usage example: use apply_chat_template for instruction-tuned model

README.md

@@ -78,18 +78,18 @@ python -c "import flash_attn; print('flash-attn OK, version:', flash_attn.__vers
 
 ## Usage
 
-Use patched Transformers as you would the standard `transformers` library — the CUDA kernels are injected transparently:
+Use patched Transformers as you would the standard `transformers` library — the CUDA kernels are injected transparently. This is an instruction-tuned model, so prompts must go through the chat template:
 
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
-model = AutoModelForCausalLM.from_pretrained("google/gemma-4-E4B-it")
+model = AutoModelForCausalLM.from_pretrained("google/gemma-4-E4B-it").cuda()
 tokenizer = AutoTokenizer.from_pretrained("google/gemma-4-E4B-it")
 
-inputs = tokenizer("Hello, how are you?", return_tensors="pt").to("cuda")
-model = model.cuda()
+messages = [{"role": "user", "content": "Hello, how are you?"}]
+inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt", return_dict=True).to("cuda")
 outputs = model.generate(**inputs, max_new_tokens=200)
-print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))
 ```
 
 ### Serving