Skip to content

Latest commit

 

History

History
49 lines (43 loc) · 1.61 KB

INFERENCE.md

File metadata and controls

49 lines (43 loc) · 1.61 KB

Demo

OpenVocab Semantic Segmentation

CUDA_VISIBLE_DEVICES=0 python inference/xdecoder/infer_semseg.py evaluate \
            --conf_files configs/xdecoder/xdecoder_focall_lang.yaml \
            --overrides \
            RESUME_FROM /pth/to/xdecoder_focall_best_openseg.pt

OpenVocab Instance Segmentation

CUDA_VISIBLE_DEVICES=0 python inference/xdecoder/infer_instseg.py evaluate \
            --conf_files configs/xdecoder/xdecoder_focall_lang.yaml \
            --overrides \
            RESUME_FROM /pth/to/xdecoder_focall_best_openseg.pt

OpenVocab Panoptic Segmentation

CUDA_VISIBLE_DEVICES=0 python inference/xdecoder/infer_panoseg.py evaluate \
            --conf_files configs/xdecoder/xdecoder_focall_lang.yaml \
            --overrides \
            RESUME_FROM /pth/to/xdecoder_focall_best_openseg.pt

OpenVocab Referring Segmentation

CUDA_VISIBLE_DEVICES=0 python inference/xdecoder/infer_refseg.py evaluate \
            --conf_files configs/xdecoder/xdecoder_focall_lang.yaml \
            --overrides \
            RESUME_FROM /pth/to/xdecoder_focall_last.pt

Region Retrieval

CUDA_VISIBLE_DEVICES=0 python inference/xdecoder/infer_region_retrieval.py evaluate \
            --conf_files configs/xdecoder/xdecoder_focall_lang.yaml \
            --overrides \
            RESUME_FROM /pth/to/xdecoder_focall_last.pt

Image Captioning

CUDA_VISIBLE_DEVICES=0 python inference/xdecoder/infer_captioning.py evaluate \
            --conf_files configs/xdecoder/xdecoder_focalt_lang.yaml \
            --overrides \
            RESUME_FROM /pth/to/xdecoder_focalt_last_novg.pt