Commit
ยท
2ca7636
1
Parent(s):
f5b3b36
Update README.md
Browse files
README.md
CHANGED
@@ -6,12 +6,14 @@ pipeline_tag: text2text-generation
|
|
6 |
---
|
7 |
|
8 |
# FLAN T5
|
|
|
|
|
9 |
FLAN T5๋ [paust/pko-t5-large](https://huggingface.co/paust/pko-t5-large) ๋ชจ๋ธ์ ๊ธฐ๋ฐ์ผ๋ก ๋ค์ํ ํ์คํฌ๋ฅผ instruction finetuning์ ํตํด์ ๋ง๋ ๋ชจ๋ธ์
๋๋ค.
|
10 |
|
11 |
ํ์ฌ ๊ณ์ Instruction Finetuning ์ ์งํํ๋ฉด์ ์ค๊ฐ๊ฒฐ๊ณผ๋ฅผ ๋ชจ๋ธ๋ก ์
๋ฐ์ดํธํ๊ณ ์์ต๋๋ค.
|
12 |
|
13 |
|
14 |
-
|
15 |
| Task name | Task type |
|
16 |
|----------------------------|----------------|
|
17 |
| NSMC | Classification |
|
@@ -36,11 +38,11 @@ FLAN T5๋ [paust/pko-t5-large](https://huggingface.co/paust/pko-t5-large) ๋ชจ
|
|
36 |
| AIHub mindslab's MRC | QA |
|
37 |
|
38 |
|
39 |
-
|
40 |
- [Hugginface ๋งํฌ](https://huggingface.co/paust/pko-flan-t5-large)
|
41 |
|
42 |
|
43 |
-
|
44 |
```python
|
45 |
from transformers import T5ForConditionalGeneration, T5TokenizerFast
|
46 |
|
@@ -53,4 +55,7 @@ input_ids = tokenizer(prompt, add_special_tokens=True, return_tensors='pt').inpu
|
|
53 |
output_ids = model.generate(input_ids=input_ids.cuda(), max_new_tokens=32, num_beams=12)
|
54 |
text = tokenizer.batch_decode(output_ids, skip_special_tokens=True)[0]
|
55 |
print(text) # ์์ธํน๋ณ์
|
56 |
-
```
|
|
|
|
|
|
|
|
6 |
---
|
7 |
|
8 |
# FLAN T5
|
9 |
+
[Source Code](https://github.com/paust/pko-t5/tree/main/pkot5/flan)
|
10 |
+
|
11 |
FLAN T5๋ [paust/pko-t5-large](https://huggingface.co/paust/pko-t5-large) ๋ชจ๋ธ์ ๊ธฐ๋ฐ์ผ๋ก ๋ค์ํ ํ์คํฌ๋ฅผ instruction finetuning์ ํตํด์ ๋ง๋ ๋ชจ๋ธ์
๋๋ค.
|
12 |
|
13 |
ํ์ฌ ๊ณ์ Instruction Finetuning ์ ์งํํ๋ฉด์ ์ค๊ฐ๊ฒฐ๊ณผ๋ฅผ ๋ชจ๋ธ๋ก ์
๋ฐ์ดํธํ๊ณ ์์ต๋๋ค.
|
14 |
|
15 |
|
16 |
+
## ํ์ต๋ ํ์คํฌ
|
17 |
| Task name | Task type |
|
18 |
|----------------------------|----------------|
|
19 |
| NSMC | Classification |
|
|
|
38 |
| AIHub mindslab's MRC | QA |
|
39 |
|
40 |
|
41 |
+
## ๋ชจ๋ธ
|
42 |
- [Hugginface ๋งํฌ](https://huggingface.co/paust/pko-flan-t5-large)
|
43 |
|
44 |
|
45 |
+
## ์ฌ์ฉ ์์
|
46 |
```python
|
47 |
from transformers import T5ForConditionalGeneration, T5TokenizerFast
|
48 |
|
|
|
55 |
output_ids = model.generate(input_ids=input_ids.cuda(), max_new_tokens=32, num_beams=12)
|
56 |
text = tokenizer.batch_decode(output_ids, skip_special_tokens=True)[0]
|
57 |
print(text) # ์์ธํน๋ณ์
|
58 |
+
```
|
59 |
+
|
60 |
+
## License
|
61 |
+
[PAUST](https://paust.io)์์ ๋ง๋ pko-t5๋ [MIT license](https://github.com/paust-team/pko-t5/blob/main/LICENSE) ํ์ ๊ณต๊ฐ๋์ด ์์ต๋๋ค.
|