arxiv:2505.20133
Konstantin Dobler
konstantindobler
·
AI & ML interests
Natural Language Processing, Transfer Learning, Crosslingual Transfer
Recent Activity
new activity about 16 hours ago
allenai/basic-skills:Convert dataset to Parquet upvoted a paper 10 months ago
AweDist: Attention-aware Embedding Distillation for New Input Token
EmbeddingsOrganizations
models 20
konstantindobler/mistral7b-ar-tokenizer-swap-pure-bf16
Text Generation • 7B • Updated • 4
konstantindobler/mistral7b-de-pure-bf16
Text Generation • 7B • Updated • 4
konstantindobler/mistral7b-de-tokenizer-swap-mixed-bf16
Text Generation • 7B • Updated • 6
konstantindobler/mistral7b-de-mixed-bf16
Text Generation • 7B • Updated • 5
konstantindobler/mistral7b-de-tokenizer-swap-pure-bf16-v2-anneal-ablation
Text Generation • 7B • Updated • 4
konstantindobler/mistral7b-de-tokenizer-swap-pure-bf16-v2
Text Generation • 7B • Updated • 4
konstantindobler/mistral7b-ar-tokenizer-swap-pure-bf16-anneal-ablation
Text Generation • 7B • Updated • 5
konstantindobler/mistral7b-de-tokenizer-swap-pure-bf16
Text Generation • 7B • Updated • 11
konstantindobler/fasttext-ar-sentencepiece-bpe-32k
Updated
konstantindobler/fasttext-de-sentencepiece-bpe-32k
Updated