 AlpinDale
|
acbdc50a71
fix: `vocab_size` field access in llava
|
hai 6 meses |
|
AlpinDale
|
e26a4ac698
chore: avoid loading the unused layers and init the VLM up to the required feature space
|
hai 6 meses |
|
AlpinDale
|
0ab35652d3
fix: llava 1.6 feature size calculation
|
hai 6 meses |
|
AlpinDale
|
7e99578712
fix: cleanup validation and update docs for vlm
|
hai 6 meses |
|
AlpinDale
|
526163003d
fix: improve consistency between feature size calc and dummy data for profiling
|
hai 6 meses |
|
AlpinDale
|
c11a8bdaad
fix: calculate max number of multi-modal tokens automatically
|
hai 6 meses |
|
AlpinDale
|
4f7d212b70
feat: remove vision language config
|
hai 6 meses |
|
AlpinDale
|
4599c98f99
feat: dynamic image size support for VLMs
|
hai 6 meses |
 AlpinDale
|
0f4a9ee77b
quantized lm_head (#582)
|
hai 6 meses |
|
AlpinDale
|
ae04f57ec1
feat: Pipeline Parallel support (#581)
|
hai 6 meses |
|
AlpinDale
|
3a0fdf7b9b
chore: remove `image_input_type` from VLM config
|
hai 6 meses |
|
AlpinDale
|
4cdc810b1c
fix: minor TP issues with vision models
|
hai 7 meses |
|
AlpinDale
|
b6ff0623a6
chore: clean up branding
|
hai 7 meses |
|
AlpinDale
|
c0c336aaa3
refactor: registry for processing model inputs; quick_gelu; clip model support
|
hai 7 meses |
|
AlpinDale
|
c5d8028668
fix: no need to redefine supports_vision and supports_lora in model class
|
hai 7 meses |
|
AlpinDale
|
56e0b8223c
chore: add base class for LoRA-supported models
|
hai 7 meses |
|
AlpinDale
|
190276865a
feat: llava-next support
|
hai 7 meses |
