google/gemma-3-27b-it
Image-Text-to-Text
β’
Updated
β’
134k
β’
624
Sungwoo Oh
sackoh
AI & ML interests
None yet
Recent Activity
new activity
3 days ago
google/gemma-3-27b-it:chat template?
liked
a model
4 days ago
google/gemma-3-27b-it
updated
a model
8 days ago
sackoh/gcp-sum-9b-8r-32a
Organizations
sackoh's activity
chat template?
1
#6 opened 3 days ago
by
apepkuss79
liked
a
model
4 days ago
updated
a
model
8 days ago
published
a
model
8 days ago
updated
a
model
8 days ago
published
a
model
8 days ago
updated
a
model
14 days ago
published
a
model
14 days ago
updated
a
model
14 days ago
published
a
model
14 days ago
updated
a
model
14 days ago
published
a
model
14 days ago
upvoted
a
paper
18 days ago
Add Optional `add_generation_prompt` Parameter to control final generation prompt in `chat_template`
1
#3 opened 4 months ago
by
sackoh
liked
a
dataset
6 months ago
liked
a
Space
8 months ago
posted
an
update
9 months ago
Post
719
π Release of open-source Korean LLM: GECKO-7B
I am delighted to share my recent project, GECKO, a bilingual large language model for Korean and English π°π·πΊπΈ. This initiative was inspired by the lack of resources for Korean large language models.
@donggyukimc and I wrote the technical report to share our insights and experiences of developing our model. While our model may not achieve sate-of-the-art performance on all benchmarks, it shows modest results with a relatively small amount of pretrained tokens.
I hope GECKO contribute to the open-source community, offering resources that can built upon and improved. I believe that through collaboration and shared knowledge, we can advance the capabilities and accessibility of large language models for Korean and other low-resource languages.
π€ Model: kifai/GECKO-7B
π Technical Report: https://arxiv.org/pdf/2405.15640
I am delighted to share my recent project, GECKO, a bilingual large language model for Korean and English π°π·πΊπΈ. This initiative was inspired by the lack of resources for Korean large language models.
@donggyukimc and I wrote the technical report to share our insights and experiences of developing our model. While our model may not achieve sate-of-the-art performance on all benchmarks, it shows modest results with a relatively small amount of pretrained tokens.
I hope GECKO contribute to the open-source community, offering resources that can built upon and improved. I believe that through collaboration and shared knowledge, we can advance the capabilities and accessibility of large language models for Korean and other low-resource languages.
π€ Model: kifai/GECKO-7B
π Technical Report: https://arxiv.org/pdf/2405.15640
- 2 replies
updated
a
model
10 months ago
Usageμ model idμ μ€νκ° μμ΅λλ€.
1
#1 opened 10 months ago
by
MrBananaHuman
authored
a
paper
10 months ago