omar-ah
/

ViL-DLM-0.6B

Image-Text-to-Text

vision-language

masked-diffusion

Model card Files Files and versions

ViL-DLM-0.6B / code

90.9 kB

Ctrl+K

Ctrl+K

1 contributor

History: 19 commits

omar-ah's picture

Add timestep-aware sparse KD weighting

25e4efd about 1 month ago

model_config.py

5.02 kB
Fix The Cauldron aokvqa config name about 1 month ago
train_production.py

59.4 kB
Add timestep-aware sparse KD weighting about 1 month ago
vil_dlm_model.py

19.3 kB
Implement stage-aware real-run training pipeline about 1 month ago
vision_xlstm.py

7.15 kB
Update model configuration and training scripts with new vision backbone support and dependencies about 1 month ago