File size: 1,186 Bytes
7bea71a
 
 
 
 
 
26e7b92
7bea71a
 
ba3f222
7bea71a
 
 
26e7b92
7bea71a
 
 
 
 
 
 
7ee1b71
7bea71a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- ko
pipeline_tag: text-generation
tags:
- finetune
license: other
---
# Model Card for mistral-ko-7b-tech
It is a fine-tuned model using Korean in the mistral-7b model.

## Model Details

* **Model Developers** :  shleeeee(Seunghyeon Lee), oopsung(Sungwoo Park)
* **Repository** : To be added
* **Model Architecture** : The mistral-ko-7b-wiki-neft is is a fine-tuned version of the Mistral-7B-v0.1.
* **Lora target modules** : q_proj, k_proj, v_proj, o_proj,gate_proj
* **train_batch** : 4
* **Max_step** : 500

## Dataset
Korean Custom Dataset(2000)

## Prompt template: Mistral
```
<s>[INST]{['instruction']}[/INST]{['output']}</s>
```

## Usage
``` 
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("shleeeee/mistral-ko-7b-tech")
model = AutoModelForCausalLM.from_pretrained("shleeeee/mistral-ko-7b-tech")

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="shleeeee/mistral-ko-7b-tech")
```

## Evaluation
![image/png](https://cdn-uploads.huggingface.co/production/uploads/654495fa893aec5da96e9134/6z75dYa8TdTy4Y7EIl0CK.png)