import torchfrom peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizer model_id = "DanishMahdi/TinyLlama-snd-sft-lora"base_id = "unsloth/tinyllama-bnb-4bit" tokenizer = AutoTokenizer.from_pretrained(model_id)model = AutoModelForCausalLM.from_pretrained( base_id, device_map="auto", torch_dtype=torch.float16,)model.resize_token_embeddings(len(tokenizer))model = PeftModel.from_pretrained(model, model_id) system = "توهان هڪ مددگار، درست ۽ محترم سنڌي AI اسسٽنٽ آهيو."user = "سنڌ جو گاديءَ وارو شهر ڪهڙو آهي؟"prompt = ( f"<|system|>\n{system}</s>\n" f"<|user|>\n{user}</s>\n" f"<|assistant|>\n")inputs = tokenizer(prompt, return_tensors="pt").to(model.device)with torch.inference_mode(): out = model.generate( **inputs, max_new_tokens=128, do_sample=True, temperature=0.7, top_p=0.9, repetition_penalty=1.05, )print(tokenizer.decode(out[0], skip_special_tokens=False))