Kevin Dave's picture

5 6 2

Kevin Dave

kxxinDave

·

AI & ML interests

Vision Models, distillation and quantisation.

Recent Activity

updated a model 16 days ago

kxxinDave/Qwen3-0.6B-ToolGRPO

updated a model 16 days ago

kxxinDave/Qwen2_5VL-JudgePO

upvoted a paper about 1 month ago

Video-Holmes: Can MLLM Think Like Holmes for Complex Video Reasoning?

View all activity

Organizations

None yet

spaces 2

Google Siglip So400m Patch14 384

SigLIP image classifier

No application file

FineTuning

models 5

kxxinDave/Qwen3-0.6B-ToolGRPO

Text Generation • 0.6B • Updated 16 days ago • 16

kxxinDave/Qwen2_5VL-JudgePO

Image-Text-to-Text • 4B • Updated 16 days ago • 36

kxxinDave/Qwen3-0.6B-GRPO-Adaptors

kxxinDave/Qwen2.5-VL-instruct-3B-Geo

Image-to-Text • 4B • Updated Mar 21 • 3 • 2

kxxinDave/Qwen2.5-VL-3B-GeoInstruct-Adapter

Image-Text-to-Text • Updated Mar 20 • 3

datasets 6

kxxinDave/GEOVQ_Qwen2_5_Geo_Description_Subset_500

Viewer • Updated Mar 21 • 500 • 4

kxxinDave/GeoSFT

Viewer • Updated Mar 19 • 8.03k • 6

kxxinDave/GEOVQ_Subset_500

Viewer • Updated Mar 12 • 500 • 7

kxxinDave/GEOVQ_Description_Subset_500

Viewer • Updated Mar 11 • 500 • 6

kxxinDave/visiondata_for_rl_hf_frompkl

Viewer • Updated Feb 10 • 441 • 6

kxxinDave/visiondata_hf_frompkl

Viewer • Updated Feb 10 • 441 • 4