LPCVC_track2/command.sh at main · zyaocoder/LPCVC_track2 · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
CUDA_VISIBLE_DEVICES=0,1 mpirun -n 2 python entry.py train \
            --conf_files configs/xdecoder/focalt_unicl_lang_finetune.yaml \
            --overrides \
            FP16 True \
            PORT 36874 \
            COCO.INPUT.IMAGE_SIZE 1024 \
            MODEL.DECODER.HIDDEN_DIM 512 \
            MODEL.ENCODER.CONVS_DIM 512 \
            MODEL.ENCODER.MASK_DIM 512 \
            MODEL.DECODER.CAPTIONING.ENABLED False \
            MODEL.DECODER.RETRIEVAL.ENABLED False \
            MODEL.DECODER.GROUNDING.ENABLED True \
            MODEL.DECODER.CAPTIONING_WEIGHT 8 \
            MODEL.DECODER.RETRIEVAL_WEIGHT 8 \
            MODEL.DECODER.TOP_CAPTIONING_LAYERS 3 \
            MODEL.DECODER.TOP_RETRIEVAL_LAYERS 3 \
            MODEL.DECODER.TOP_GROUNDING_LAYERS 6 \
            MODEL.DECODER.GROUNDING.TEXT_WEIGHT 2.0 \
            MODEL.DECODER.GROUNDING.CLASS_WEIGHT 0.5 \
            COCO.TEST.BATCH_SIZE_TOTAL 4 \
            COCO.TRAIN.BATCH_SIZE_TOTAL 4 \
            COCO.TRAIN.BATCH_SIZE_PER_GPU 2 \
            VLP.TEST.BATCH_SIZE_TOTAL 2 \
            VLP.TRAIN.BATCH_SIZE_TOTAL 16 \
            VLP.TRAIN.BATCH_SIZE_PER_GPU 8 \
            VLP.DATALOADER.NUM_WORKERS 16 \
            ADE20K.TEST.BATCH_SIZE_TOTAL 2 \
            REF.TEST.BATCH_SIZE_TOTAL 2 \
            SOLVER.LR_MULTIPLIER.lang_encoder 0.1 \
            WEIGHT True \
            RESUME_FROM pretrained_weights/focalt_in21k_yfcc_gcc_xdecoder_unicl.pt


# CUDA_VISIBLE_DEVICES=1 mpirun -n 1 python entry.py evaluate \
#             --conf_files configs/xdecoder/focalt_unicl_lang_finetune.yaml \
#             --overrides \
#             COCO.INPUT.IMAGE_SIZE 1024 \
#             MODEL.DECODER.CAPTIONING.ENABLED False \
#             MODEL.DECODER.RETRIEVAL.ENABLED False \
#             MODEL.DECODER.GROUNDING.ENABLED True \
#             COCO.TEST.BATCH_SIZE_TOTAL 1 \
#             COCO.TRAIN.BATCH_SIZE_TOTAL 1 \
#             COCO.TRAIN.BATCH_SIZE_PER_GPU 1 \
#             VLP.TEST.BATCH_SIZE_TOTAL 1 \
#             VLP.TRAIN.BATCH_SIZE_TOTAL 1 \
#             VLP.TRAIN.BATCH_SIZE_PER_GPU 1 \
#             MODEL.DECODER.HIDDEN_DIM 512 \
#             MODEL.ENCODER.CONVS_DIM 512 \
#             MODEL.ENCODER.MASK_DIM 512 \
#             ADE20K.TEST.BATCH_SIZE_TOTAL 1 \
#             FP16 True \
#             WEIGHT True \
#             RESUME_FROM /pth/to/xdecoder_data/pretrained/focalt_in21k_yfcc_gcc_xdecoder_unicl.pt