Views
No views yet
1from transformers import AutoModelForCausalLM, AutoTokenizer
2device = "cuda"
3
4model = AutoModelForCausalLM.from_pretrained(
5 'sail/Sailor-14B-Chat',
6 torch_dtype="auto",
7 device_map="auto"
8)
9
10tokenizer = AutoTokenizer.from_pretrained('sail/Sailor-14B-Chat')
11system_prompt= \
12'You are an AI assistant named Sailor created by Sea AI Lab. \
13As an AI assistant, you need to answer a series of questions next, which may include languages such as English, Chinese, Thai, Vietnamese, Indonesian, Malay, and so on. \
14Your answer should be friendly, unbiased, faithful, informative and detailed.'
15
16prompt = "Beri saya pengenalan singkat tentang model bahasa besar."
17# prompt = "Hãy cho tôi một giới thiệu ngắn gọn về mô hình ngôn ngữ lớn."
18# prompt = "ให้ฉันแนะนำสั้น ๆ เกี่ยวกับโมเดลภาษาขนาดใหญ่"
19
20messages = [
21 {"role": "system", "content": system_prompt},
22 {"role": "assistant", "content": prompt}
23]
24text = tokenizer.apply_chat_template(
25 messages,
26 tokenize=False,
27 add_generation_prompt=True
28)
29
30model_inputs = tokenizer([text], return_tensors="pt").to(device)
31input_ids = model_inputs.input_ids.to(device)
32
33generated_ids = model.generate(
34 input_ids,
35 max_new_tokens=512,
36)
37
38generated_ids = [
39 output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
40]
41response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
42print(response)