8000 History for gliner2/training/lora.py - kiminh/GLiNER2 · GitHub

Repository navigation

Commits

History for

Commits on Oct 8, 2026
- Merge pull request #160 from jaanavit/fix/unload-adapter-noop-on-non-peftmodel
  urchade
  authored
  d30035a
Commits on Sep 23, 2026
- Import peft only on LoRA code paths so non-LoRA training works without it
  
  hfastino
  and
  cursoragent
  committed
  68f0d4b
Commits on Sep 22, 2026
- Fix unload_lora_adapter() silent no-op when model is not a PeftModel
  
  jaanavit
  and
  cursoragent
  committed
  0a44ba0
Commits on Aug 17, 2026
- add structured prediction support to g2.5
  urchade
  committed
  d407a2e
Commits on Jul 27, 2026
- add boundary architecture
  urchade
  committed
  90b0dd8
Commits on Apr 21, 2026
- ENG-1268: refactor LoRA internals to PEFT backend
  MattThomas-fastino
  committed
  03c722f
Commits on Dec 18, 2025
- fix wandb logging and lora to gpu
  urchade
  committed
  20d03b5
Commits on Dec 17, 2025
- fix wandb logging and lora to gpu
  urchade
  committed
  f7126a3
Commits on Dec 16, 2025
- smol fix
  urchade
  committed
  a378f10
- add more customization for lora
  urchade
  committed
  1136cab
Commits on Dec 12, 2025
- add lora adapter
  urchade
  committed
  791bdaf
- fix weight access for lora
  urchade
  committed
  a0159ee
- fix lora param counting bug
  urchade
  committed
  cf7a611
Commits on Dec 10, 2025
- add lora
  urchade
  committed
  3dd2380
Loading

0