jinaai
/

jina-bert-flash-implementation

🇪🇺 Region: EU

Model card Files Files and versions

jina-bert-flash-implementation

Ctrl+K

Ctrl+K

6 contributors

History: 108 commits

Markus28's picture

Set activation_checkpoint_lvl to 100 by default

535ad9a about 2 years ago

README.md

1.89 kB
feat: added README about 2 years ago
bert_padding.py

9.78 kB
reference the flash attention GitHub about 2 years ago
block.py

17.4 kB
reference the flash attention GitHub about 2 years ago
configuration_bert.py

7.33 kB
Set activation_checkpoint_lvl to 100 by default about 2 years ago
convert_v2_weights.py

6.1 kB
feat: for converting v2, added lines to save model weights and print config about 2 years ago
embedding.py

2.26 kB
clean up embeddings.py (#7) about 2 years ago
mha.py

35.3 kB
reference the flash attention GitHub about 2 years ago
mlp.py

7.79 kB
Porting v2 models to flash attention (#15) about 2 years ago
modeling_bert.py

33.9 kB
feat: added selective activation checkpointing about 2 years ago
modeling_for_glue.py

10.7 kB
feat: assert return_dict about 2 years ago
modeling_lora.py

12.3 kB
fix: use staticmethod istead of classmethod about 2 years ago