eenzeenee commited on
Commit
442879a
โ€ข
1 Parent(s): fc3331d

Update README.md

Browse files

add training args

Files changed (1) hide show
  1. README.md +17 -0
README.md CHANGED
@@ -9,6 +9,7 @@ tags:
9
  # t5-base-korean-summarization
10
 
11
  This is [T5](https://huggingface.co/docs/transformers/model_doc/t5) model for korean text summarization.
 
12
  Finetuned with 3 datasets. Specifically, it is described below.
13
 
14
  - [Korean Paper Summarization Dataset(๋…ผ๋ฌธ์ž๋ฃŒ ์š”์•ฝ)](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=realm&dataSetSn=90)
@@ -81,6 +82,22 @@ RESULT >> ๊ตญ์–ด ์—ฌํ–‰์„ ๋– ๋‚˜๊ธฐ ์ „์— ๊ตญ์–ด ์—ฌํ–‰์„ ๋– ๋‚  ์ค€๋น„๋ฌผ๊ณผ
81
 
82
  # Training
83
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
84
  # Model Architecture
85
 
86
  ```
 
9
  # t5-base-korean-summarization
10
 
11
  This is [T5](https://huggingface.co/docs/transformers/model_doc/t5) model for korean text summarization.
12
+ Finetuned based on ['paust/pko-t5-base'](https://huggingface.co/paust/pko-t5-base) model.
13
  Finetuned with 3 datasets. Specifically, it is described below.
14
 
15
  - [Korean Paper Summarization Dataset(๋…ผ๋ฌธ์ž๋ฃŒ ์š”์•ฝ)](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=realm&dataSetSn=90)
 
82
 
83
  # Training
84
 
85
+ The model was trained with the parameters:
86
+
87
+ ### training arguments
88
+ ```
89
+ Seq2SeqTrainingArguments(
90
+ per_device_train_batch_size=8,
91
+ per_device_eval_batch_size=8,
92
+ auto_find_batch_size=False,
93
+ weight_decay=0.01,
94
+ learning_rate=4e-05,
95
+ lr_scheduler_type=linear,
96
+ num_train_epochs=3,
97
+ fp16=True)
98
+ ```
99
+
100
+
101
  # Model Architecture
102
 
103
  ```