jinaai / jina-bert-flash-implementation

Last scanned: Never Scanned

Unscanned

Versions

Version

Commit message

Commit author

Last scanned

feat-add-configs (#18)koukandreMar 21, 2025
fix: handle window_size passed as listgmastrapasOct 21, 2024
fix-glu-mlp (#17)michael-guentherOct 21, 2024
Porting v2 models to flash attention (#15)bwang0911Oct 21, 2024
feat: updated activation checkpointing (#14)Markus28Oct 21, 2024
fix: use staticmethod istead of classmethodMarkus28Oct 21, 2024
feat: added docstringsMarkus28Oct 19, 2024
feat: Allow LoRA to be merged into weights (#12)Markus28Oct 21, 2024
removed num_tasks from configMarkus28Nov 11, 2024
fix: remove cleaving (#13)Markus28Nov 11, 2024
fix: added trust_remote_code to tokenizerMarkus28Nov 11, 2024
fix: fixed from_bert methodMarkus28Nov 11, 2024
fix: fix LoRA implementationMarkus28Nov 8, 2024
feat: cleave off layers from encoder (#11)Markus28Nov 8, 2024
feat: only apply select_task_for_layer if task has changedMarkus28Nov 11, 2024
feat: make num of loras part of the configMarkus28Nov 11, 2024
feat: added separate BertForMaskedLM classMarkus28Nov 11, 2024
feat: added return_dictMarkus28Oct 21, 2024
feat: add current_task to forwardMarkus28Oct 22, 2024
feat: use property in LoRA parametrizationMarkus28Oct 23, 2024
Protect AI's security scanner detects threats in model files
With Protect AI's Guardian you can scan models for threats before ML developers download them for use, and apply policies based on your risk tolerance.
Learn more
Found a new threat?
Earn bounties and get recognition for your discoveries by submitting them through Huntr.
Report your finding