comhu0503 commited on
Commit
052d420
·
verified ·
1 Parent(s): 8fe0a4a

Add minimal model card (no training data)

Browse files
Files changed (1) hide show
  1. README.md +28 -0
README.md ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ base_model: skt/A.X-4.0-Light
4
+ pipeline_tag: text-generation
5
+ language:
6
+ - ko
7
+ ---
8
+
9
+ # axlight-4.0-light-merged
10
+
11
+ `skt/A.X-4.0-Light` 를 도메인 적응 파인튜닝한 한국어 생성 모델입니다(LoRA 병합본).
12
+ 단독으로 바로 추론에 사용할 수 있습니다.
13
+
14
+ ## 사용법
15
+
16
+ ```python
17
+ from transformers import AutoModelForCausalLM, AutoTokenizer
18
+ import torch
19
+
20
+ tok = AutoTokenizer.from_pretrained("comhu/axlight-4.0-light-merged")
21
+ model = AutoModelForCausalLM.from_pretrained(
22
+ "comhu/axlight-4.0-light-merged", torch_dtype=torch.bfloat16, device_map="auto")
23
+
24
+ msgs = [{"role": "user", "content": "회사 소개를 한 문단으로 작성해줘."}]
25
+ ids = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt").to(model.device)
26
+ out = model.generate(ids, max_new_tokens=512, do_sample=False)
27
+ print(tok.decode(out[0][len(ids[0]):], skip_special_tokens=True))
28
+ ```