File size: 1,084 Bytes
86b54c5
 
 
 
4ae37c0
 
 
86b54c5
e0a32dd
efd6d72
e0a32dd
 
f482b55
 
efd6d72
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
tags:
- lightblue/karasu-1.1B
- yuiseki/karasu-sake-qa-v0.1
language:
- ja
- en
---
# karasu-lora-jp-qa-chat

karasu fine tuned model by lora method with the original Q&A dataset.
This model is usefull for rag system with reference input files.
* Base model: Merged model of lightblue/karasu-1.1B and yuiseki/karasu-sake-qa-v0.1 (aipib/karasu-slerp1)
* Traning dataset: Original japanese question and answer lists

## 💻 Usage

```python
!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "aipib/suzume-taskarith1"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])
```