File size: 1,243 Bytes
322906a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 |
---
license: apache-2.0
language:
- ko
pipeline_tag: text-generation
---
## Prompt Tempalte
It follows Alpaca format.
```
### 질문: {instruction}
### 답변: {output}
```
### Implementation Code
```
import troch
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.fron_pretrained("Ja3ck/Mistral-instruct-IPO-Y24-v1", return_dict=True, torch_dtype=torch.bfloat16, device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("Ja3ck/Mistral-instruct-IPO-Y24-v1", use_fast=True)
tokenizer.pad_token = tokenizer.unk_token
tokenizer.pad_token_id = tokenizer.unk_token_id
tokenizer.padding_side = "left"
def gen(x):
x_ = f"### 질문: {x.strip()} ### 답변: "
inputs = tokenizer(x_, return_tensor='pt')
input_ids = inputs['input_ids'].cuda()
generation_output = model.generate(
pad_token_id = tokenizer.pad_token_id,
temperature=0.1,
top_p=1,
top_k=50,
num_beams=1,
repetition_penalty=1.13,
do_sample=True,
),
return_dict_in_generate=True,
output_scores=True,
max_new_tokens=1024
)
for seq in generation_output.sequences:
output = tokenizer.decode(seq)
print(output.split("### 답변: ")[1].strip())
gen("안녕하세요?")
``` |