-
VTM-1.5.1.pt = the release checkpoint (10-step teacher, 1 or 2 steps); card: how it works, download chart, commercial-use column, license: other
-
Point the card at the renamed repo sinBoo1/VTM-Spark
-
Update model card: VTM-1.5.1, VTM Spark, demo, files and licences
-
Add trackers/animeseg_hair3.pt
-
Upload VTM-1.5.1.pt
latest model it adds SFT and AD at 7k steps and 6k steps -
Delete VTM-1.5.1.pt
-
Delete VTM-ELF.pt
-
Upload VTM-1.5.1.pt
-
Add openseeface/priorbox_640x640.json
-
Add openseeface/mnv3_gaze32_split_opt.onnx
-
Add openseeface/mnv3_detection_opt.onnx
-
Add openseeface/retinaface_640x640_opt.onnx
-
Add openseeface/lm_model3_opt.onnx
-
Add trackers/pose_landmarker_lite.task
-
Add trackers/dwpose_v2.pt
-
Add trackers/iris_pose.pt
-
Upload VTM-ELF.pt
-
Delete VTM-elf.pt-000060000
-
Delete method illustration
-
Remove method illustration and caption
-
Update method figure: detailed face keypoints (brows, 3-pt eyes, mouth)
-
Simplify card: torso limits, shorter architecture, new illustration
-
Revise model card: architecture, 7 FPS note, method figure
-
Expand model card: VTuber pose-transfer beta docs
-
Update README.md
-
Upload VTM-elf.pt-000060000
base model -
initial commit
