Youngjoon Jang

yjoonjang

·

https://yjoonjang.github.io/

AI & ML interests

Information Retrieval (IR), Retrieval-Augmented Generation (RAG)

Recent Activity

updated a collection about 14 hours ago

Retrieval models I've trained

updated a collection about 14 hours ago

Retrieval models I've trained

updated a collection about 14 hours ago

View all activity

Organizations

New activity in yjoonjang/colbert-ko-v1 13 days ago

Add Sentence Transformers usage

#2 opened 15 days ago by

New activity in whybe-choi/ko-vdr-hn about 2 months ago

Update README with old README

#15 opened about 2 months ago by

Restore dataset_info for base configs (corpus/naive/marginpos/percpos/shifted_by_n)

#14 opened about 2 months ago by

Add distill_hardest + ensembledistill_{stratified,hardest} subsets

#13 opened about 2 months ago by

Add ensembledistill_hardest subset

#12 opened about 2 months ago by

Add ensembledistill_stratified subset

#11 opened about 2 months ago by

Add distill_hardest subset

#10 opened about 2 months ago by

Add distill_stratified subset

#9 opened about 2 months ago by

Add reranked_hardest config & rename reranked -> reranked_stratified

#8 opened about 2 months ago by

Add ko-vdr-train-public-v2.0_reranked config (KL-distillation set)

#7 opened about 2 months ago by

Add ko-vdr-train-public-v2.0_reranked config (KL-distillation set)

#6 opened about 2 months ago by

New activity in whybe-choi/ko-vdr-hn 2 months ago

Replace ko-vdr-train-public-v2.0_reranked with training-ready flat schema (positive + 8 negatives + reranked_scores)

#5 opened 2 months ago by

Rename reranker config to ko-vdr-train-public-v2.0_reranked + note source coverage

#4 opened 2 months ago by

Add ko-vdr-train-public-v2.0_reranked config (Qwen3-VL reranker-scored candidates, mined from ko-vdr-train-public-v2.0 only)

#3 opened 2 months ago by

Add reranked_qwen3_vl_8b config (Qwen3-VL reranker-scored candidates for distillation)

#2 opened 2 months ago by

New activity in google/embeddinggemma-300m 7 months ago

Question regarding the two-layer linear (FNN) projection structure in EmbeddingGemma-300M

#33 opened 9 months ago by

New activity in yjoonjang/inference-free-splade-ko-v1 7 months ago

안녕하세요.

#1 opened 7 months ago by

New activity in skt/A.X-Encoder-base 7 months ago

Set tokenizer "model_max_length" property to 16384

#1 opened 7 months ago by

New activity in yjoonjang/colbert-ko-v1 10 months ago

Produce ColBERT-KO Evaluation Results

#1 opened 11 months ago by

New activity in lightonai/embeddings_supervised 11 months ago

How are the negatives mined?

#2 opened 11 months ago by