|
--- |
|
language: |
|
- ko |
|
- en |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- >- |
|
KETI-AIR/aihub_koenzh_food_translation,KETI-AIR/aihub_scitech_translation,KETI-AIR/aihub_scitech20_translation,KETI-AIR/aihub_socialtech20_translation,KETI-AIR/aihub_spoken_language_translation |
|
metrics: |
|
- bleu |
|
model-index: |
|
- name: ko2en_bidirection2 |
|
results: |
|
- task: |
|
name: Translation |
|
type: translation |
|
dataset: |
|
name: >- |
|
KETI-AIR/aihub_koenzh_food_translation,KETI-AIR/aihub_scitech_translation,KETI-AIR/aihub_scitech20_translation,KETI-AIR/aihub_socialtech20_translation,KETI-AIR/aihub_spoken_language_translation |
|
koen,none,none,none,none |
|
type: >- |
|
KETI-AIR/aihub_koenzh_food_translation,KETI-AIR/aihub_scitech_translation,KETI-AIR/aihub_scitech20_translation,KETI-AIR/aihub_socialtech20_translation,KETI-AIR/aihub_spoken_language_translation |
|
args: koen,none,none,none,none |
|
metrics: |
|
- name: Bleu |
|
type: bleu |
|
value: 51.5949 |
|
license: apache-2.0 |
|
pipeline_tag: translation |
|
widget: |
|
- text: "translate_ko2en: IBM 왓슨X는 AI 및 데이터 플랫폼이다. 신뢰할 수 있는 데이터, 속도, 거버넌스를 갖고 파운데이션 모델 및 머신 러닝 기능을 포함한 AI 모델을 학습시키고, 조정해, 조직 전체에서 활용하기 위한 전 과정을 아우르는 기술과 서비스를 제공한다." |
|
example_title: "KO2EN 1" |
|
- text: "translate_ko2en: 이용자는 신뢰할 수 있고 개방된 환경에서 자신의 데이터에 대해 자체적인 AI를 구축하거나, 시장에 출시된 AI 모델을 정교하게 조정할 수 있다. 대규모로 활용하기 위한 도구 세트, 기술, 인프라 및 전문 컨설팅 서비스를 활용할 수 있다." |
|
example_title: "KO2EN 2" |
|
- text: "translate_en2ko: The Seoul Metropolitan Government said Wednesday that it would develop an AI-based congestion monitoring system to provide better information to passengers about crowd density at each subway station." |
|
example_title: "EN2KO 1" |
|
- text: "translate_en2ko: According to Seoul Metro, the operator of the subway service in Seoul, the new service will help analyze the real-time flow of passengers and crowd levels in subway compartments, improving operational efficiency." |
|
example_title: "EN2KO 2" |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# ko2en_bidirection2 |
|
|
|
This model is a fine-tuned version of [KETI-AIR/long-ke-t5-base](https://huggingface.co/KETI-AIR/long-ke-t5-base) on the KETI-AIR/aihub_koenzh_food_translation,KETI-AIR/aihub_scitech_translation,KETI-AIR/aihub_scitech20_translation,KETI-AIR/aihub_socialtech20_translation,KETI-AIR/aihub_spoken_language_translation koen,none,none,none,none dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.5716 |
|
- Bleu: 51.5949 |
|
- Gen Len: 28.8348 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 0.001 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 8 |
|
- total_train_batch_size: 128 |
|
- total_eval_batch_size: 128 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 3.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |
|
|:-------------:|:-----:|:------:|:---------------:|:-------:|:-------:| |
|
| 0.7004 | 1.0 | 187524 | 0.6461 | 28.0622 | 17.8368 | |
|
| 0.6176 | 2.0 | 375048 | 0.5967 | 29.3033 | 17.8281 | |
|
| 0.5642 | 3.0 | 562572 | 0.5716 | 30.0045 | 17.8366 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.25.1 |
|
- Pytorch 1.12.0 |
|
- Datasets 2.8.0 |
|
- Tokenizers 0.13.2 |
