Add model card explaining what these weights are and where they came from

feat: add pwtf-dvd face alignment weights

feat: add clip-vit-base-patch16 for Yermandy

feat: add ext_cache (CLIP, SigLIP, SD VAE, EfficientNet, VideoSeal, AudioSeal)

sync from GitHub (e4aeaa8)

sync from GitHub (0154d02)

chore: remove junk files (macOS artifacts, zips, gitattributes)

chore: remove non-weight files (batch 7)

chore: remove non-weight files (batch 6)

chore: remove non-weight files (batch 5)

chore: remove non-weight files (batch 4)

chore: remove non-weight files (batch 3)

chore: remove non-weight files (batch 2)

chore: remove non-weight files (batch 1)

Upload video/fake-stormer/weights/fakestormer_c23.pth with huggingface_hub

Upload video/fake-stormer/weights/TopDownDetector_C23_ViTBase224_ST_hm100_tempLOC0.2_4SBI_SAM_mp0.01_temp2_0vidAug_0.35_temp_normlms_normREAL_model_best.pth with huggingface_hub

Upload video/fake-stormer/weights/TopDownDetector_C0_ViTBase224_ST_hm100_tempLOC0.2_4SBI_SAM_mp0.01_temp2_0vidAug_0.25_temp_normlms_IN_model_best.pth with huggingface_hub

Upload video/dfd-fcg/weights/dfd_fcg_checkpoint.pth with huggingface_hub

Upload audio/sonics/weights/pytorch_model.bin with huggingface_hub

Upload audio/sonics/weights/config.json with huggingface_hub

Upload audio/sonics/weights/README.md with huggingface_hub

Upload image/yermandy/weights/model.ckpt with huggingface_hub

Upload image/universal/weights/fc_weights.pth with huggingface_hub

Upload image/npr/weights/NPR.pth with huggingface_hub

sync from GitHub (81cf3c3)

sync from GitHub (18c28bb)

sync from GitHub (5cbbb56)

sync from GitHub (bb0bd84)

sync from GitHub (c4c7b2c)

sync from GitHub (17382d4)

sync from GitHub (28464ab)

sync from GitHub (7713d2c)

sync from GitHub (3150c97)

feat: add 5 new video detection services (DFD-FCG, PwTF-DVD, LipFD, RECCE, MINTIME)

feat: GPU enablement -- 3-tier CUDA Dockerfiles, TF32, health reporting (2026-04-02)

Upload DeepSafe services: model weights and inference code

initial commit