Markus Krimmel

Markus28's activity

• All
• Models
• Datasets
• Spaces
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in jinaai/jina-bert-flash-implementation almost 2 years ago

Porting v2 models to flash attention

🔥 1

#15 opened almost 2 years ago by

Markus28

feat: updated activation checkpointing

#14 opened almost 2 years ago by

Markus28

feat: Allow LoRA to be merged into weights

#12 opened almost 2 years ago by

Markus28

fix: remove cleaving

#13 opened almost 2 years ago by

Markus28

feat: cleave off layers from encoder

#11 opened almost 2 years ago by

Markus28

clean up embeddings.py

#7 opened almost 2 years ago by

bwang0911

support-multiple-task-ids

#5 opened almost 2 years ago by

michael-guenther

feat: implement task type embeddings

#1 opened almost 2 years ago by

Markus28

New activity in jinaai/jina-bert-v2-qk-devlin-norm-1e-2 almost 2 years ago

Use attention dropout during training

1

#1 opened almost 2 years ago by

Markus28

New activity in jinaai/jina-bert-implementation almost 2 years ago

Use attention dropout during training

2

#10 opened almost 2 years ago by

Markus28

New activity in jinaai/jina-bert-implementation about 2 years ago

Fix sorting heuristic

1

#3 opened about 2 years ago by

Markus28