Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

About the segmentation and detection. I use the imagenet-1k pretrained model to segmentation and detection task. but the mIoU I get is very low #88

Open
1104662797 opened this issue Dec 23, 2021 · 4 comments
Open

About the segmentation and detection. I use the imagenet-1k pretrained model to segmentation and detection task. but the mIoU I get is very low #88

1104662797 opened this issue Dec 23, 2021 · 4 comments

Comments

@1104662797
Copy link

I use pvt_tiny for this two jobs. But the mIoU is very low. we train with this script on two rtx3090, I just change the gpu_multiples to 1:

_base_ = [
    '../../_base_/models/fpn_r50.py',
    '../../_base_/datasets/ade20k.py',
    '../../_base_/default_runtime.py'
]
# model settings
model = dict(
    type='EncoderDecoder',
    # pretrained='pretrained/pvt_tiny.pth',
    pretrained='/home/host/dist1/zx/classification/checkpoints/pvt_tiny/checkpoint.pth',
    backbone=dict(
        type='pvt_tiny',
        style='pytorch'),
    neck=dict(in_channels=[64, 128, 320, 512]),
    decode_head=dict(num_classes=150))


gpu_multiples = 1  # we use 8 gpu instead of 4 in mmsegmentation, so lr*2 and max_iters/2
# optimizeiiiiiir
optimizer = dict(type='AdamW', lr=0.0001*gpu_multiples, weight_decay=0.0001)
optimizer_config = dict()
# learning policy
lr_config = dict(policy='poly', power=0.9, min_lr=0.0, by_epoch=False)
# runtime settings
runner = dict(type='IterBasedRunner', max_iters=80000//gpu_multiples)
checkpoint_config = dict(by_epoch=False, interval=8000//gpu_multiples)
evaluation = dict(interval=8000//gpu_multiples, metric='mIoU')

the part of log I get :

       Class        |  IoU  |  Acc  |
+---------------------+-------+-------+
|         wall        |  54.7 | 81.83 |
|       building      | 67.78 | 89.28 |
|         sky         |  89.7 | 94.46 |
|        floor        | 57.27 | 73.54 |
|         tree        | 58.49 | 78.73 |
|       ceiling       | 63.11 | 73.05 |
|         road        | 64.98 | 79.87 |
|         bed         | 48.05 | 77.07 |
|      windowpane     | 40.18 | 60.33 |
|        grass        | 57.12 | 75.26 |
|       cabinet       | 32.02 |  46.5 |
|       sidewalk      | 41.35 |  63.2 |
|        person       | 40.43 | 64.08 |
|        earth        | 27.29 | 36.52 |
|         door        |  5.95 |  6.61 |
|        table        | 18.18 | 30.23 |
|       mountain      | 31.44 | 51.38 |
|        plant        | 33.12 | 43.71 |
|       curtain       | 38.04 | 52.99 |
|        chair        | 21.51 | 36.51 |
|         car         |  51.0 | 71.94 |
|        water        | 29.28 | 36.67 |
|       painting      | 31.18 | 51.95 |
|         sofa        | 24.42 | 43.53 |
|        shelf        | 13.72 | 23.72 |
|        house        | 28.35 | 42.56 |
|         sea         | 42.37 | 68.75 |
|        mirror       | 11.43 |  14.2 |
|         rug         | 23.17 | 27.49 |
|        field        | 22.71 | 47.06 |
|       armchair      |  0.78 |  0.84 |
|         seat        | 25.22 |  31.6 |
|        fence        |  4.03 |  4.5  |
|         desk        | 10.89 |  15.5 |
|         rock        |  16.8 | 28.19 |
|       wardrobe      |  7.48 |  7.94 |
|         lamp        | 14.81 |  19.2 |
|       bathtub       | 13.31 | 16.23 |
|       railing       |  5.92 |  6.01 |
|       cushion       | 13.75 | 20.18 |
|         base        |  0.55 |  0.62 |
|         box         |  1.56 |  1.72 |
|        column       |  0.0  |  0.0  |
|      signboard      |  4.54 |  4.79 |
|   chest of drawers  | 12.02 | 17.29 |
|       counter       |  0.69 |  0.7  |
|         sand        | 14.07 | 21.83 |
|         sink        | 20.21 | 35.94 |
|      skyscraper     | 31.83 | 37.65 |
|      fireplace      | 26.12 | 43.07 |
|     refrigerator    | 18.92 | 23.57 |
|      grandstand     | 21.16 | 30.05 |
|         path        | 10.79 | 12.48 |
|        stairs       | 10.18 | 11.39 |
|        runway       | 48.93 | 73.64 |
|         case        |  24.5 | 49.33 |
|      pool table     | 50.19 | 63.41 |
|        pillow       | 18.73 | 23.37 |
|     screen door     | 17.06 | 19.77 |
|       stairway      |  5.3  |  7.01 |
|        river        |  9.48 | 16.44 |
|        bridge       |  0.28 |  0.29 |
|       bookcase      |  6.67 |  8.51 |
|        blind        |  1.8  |  1.84 |
|     coffee table    | 14.99 | 27.61 |
|        toilet       | 21.48 | 47.07 |
|        flower       | 14.28 |  23.6 |
|         book        | 15.55 | 18.32 |
|         hill        |  3.46 |  3.74 |
|        bench        |  5.57 |  5.82 |
|      countertop     |  7.53 |  9.42 |
|        stove        | 18.68 | 36.78 |
|         palm        | 10.27 | 11.79 |
|    kitchen island   | 15.96 |  18.7 |
|       computer      |  6.35 |  8.5  |
|     swivel chair    | 12.05 |  20.8 |
|         boat        |  5.43 |  6.27 |
|         bar         |  0.04 |  0.04 |
|    arcade machine   |  5.5  |  6.23 |
|        hovel        |  3.51 |  4.12 |
|         bus         |  6.68 |  7.19 |
|        towel        |  2.35 |  2.37 |
|        light        | 10.98 | 11.26 |
|        truck        |  0.1  |  0.13 |
|        tower        | 10.42 | 10.66 |
|      chandelier     | 29.94 | 50.31 |
|        awning       |  0.35 |  0.38 |
|     streetlight     |  0.28 |  0.28 |
|        booth        |  0.0  |  0.0  |
| television receiver | 17.11 | 24.19 |
|       airplane      | 31.53 | 39.65 |
|      dirt track     |  0.0  |  0.0  |
|       apparel       |  4.78 |  5.66 |
|         pole        |  1.96 |  2.17 |
|         land        |  0.0  |  0.0  |
|      bannister      |  0.0  |  0.0  |
|      escalator      |  0.0  |  0.0  |
|       ottoman       |  0.0  |  0.0  |
|        bottle       |  0.0  |  0.0  |
|        buffet       |  0.0  |  0.0  |
|        poster       |  0.0  |  0.0  |
|        stage        |  0.0  |  0.0  |
|         van         |  0.4  |  0.4  |
|         ship        |  0.0  |  0.0  |
|       fountain      |  0.0  |  0.0  |
|    conveyer belt    |  0.0  |  0.0  |
|        canopy       |  0.0  |  0.0  |
|        washer       |  0.28 |  0.28 |
|      plaything      |  0.5  |  0.54 |
|    swimming pool    | 24.38 | 42.56 |
|        stool        |  0.0  |  0.0  |
|        barrel       |  0.0  |  0.0  |
|        basket       |  0.0  |  0.0  |
|      waterfall      | 47.73 | 70.95 |
|         tent        | 31.71 | 47.17 |
|         bag         |  0.0  |  0.0  |
|       minibike      | 10.11 | 11.68 |
|        cradle       | 29.79 | 67.07 |
|         oven        |  0.0  |  0.0  |
|         ball        |  0.17 |  0.38 |
|         food        | 27.43 | 37.67 |
|         step        |  0.0  |  0.0  |
|         tank        |  0.0  |  0.0  |
|      trade name     |  3.35 |  3.46 |
|      microwave      | 10.61 | 12.33 |
|         pot         |  0.0  |  0.0  |
|        animal       |  1.39 |  1.51 |
|       bicycle       |  0.0  |  0.0  |
|         lake        |  2.98 |  3.54 |
|      dishwasher     |  2.75 |  2.79 |
|        screen       | 21.23 | 22.22 |
|       blanket       |  0.0  |  0.0  |
|      sculpture      |  0.0  |  0.0  |
|         hood        |  3.97 |  5.03 |
|        sconce       |  1.98 |  2.02 |
|         vase        |  2.1  |  2.49 |
|    traffic light    |  0.03 |  0.03 |
|         tray        |  0.0  |  0.0  |
|        ashcan       |  0.0  |  0.0  |
|         fan         | 12.51 | 14.54 |
|         pier        |  0.68 |  0.68 |
|      crt screen     |  0.0  |  0.0  |
|        plate        |  0.0  |  0.0  |
|       monitor       |  0.0  |  0.0  |
|    bulletin board   |  0.0  |  0.0  |
|        shower       |  0.0  |  0.0  |
|       radiator      |  0.0  |  0.0  |
|        glass        |  0.0  |  0.0  |
|        clock        |  0.0  |  0.0  |
|         flag        |  0.0  |  0.0  |
+---------------------+-------+-------+
2021-12-23 16:21:00,788 - mmseg - INFO - Summary:
2021-12-23 16:21:00,789 - mmseg - INFO - 
+-------+-------+-------+
|  aAcc |  mIoU |  mAcc |
+-------+-------+-------+
| 65.06 | 14.33 | 20.34 |
+-------+-------+-------+
2021-12-23 16:21:00,804 - mmseg - INFO - Exp name: fpn_pvt_t_ade20k_40k.py
2021-12-23 16:21:00,805 - mmseg - INFO - Iter(val) [1000]	aAcc: 0.6506, mIoU: 0.1433, mAcc: 0.2034, IoU.wall: 0.5470, IoU.building: 0.6778, IoU.sky: 0.8970, IoU.floor: 0.5727, IoU.tree: 0.5849, IoU.ceiling: 0.6311, IoU.road: 0.6498, IoU.bed : 0.4805, IoU.windowpane: 0.4018, IoU.grass: 0.5712, IoU.cabinet: 0.3202, IoU.sidewalk: 0.4135, IoU.person: 0.4043, IoU.earth: 0.2729, IoU.door: 0.0595, IoU.table: 0.1818, IoU.mountain: 0.3144, IoU.plant: 0.3312, IoU.curtain: 0.3804, IoU.chair: 0.2151, IoU.car: 0.5100, IoU.water: 0.2928, IoU.painting: 0.3118, IoU.sofa: 0.2442, IoU.shelf: 0.1372, IoU.house: 0.2835, IoU.sea: 0.4237, IoU.mirror: 0.1143, IoU.rug: 0.2317, IoU.field: 0.2271, IoU.armchair: 0.0078, IoU.seat: 0.2522, IoU.fence: 0.0403, IoU.desk: 0.1089, IoU.rock: 0.1680, IoU.wardrobe: 0.0748, IoU.lamp: 0.1481, IoU.bathtub: 0.1331, IoU.railing: 0.0592, IoU.cushion: 0.1375, IoU.base: 0.0055, IoU.box: 0.0156, IoU.column: 0.0000, IoU.signboard: 0.0454, IoU.chest of drawers: 0.1202, IoU.counter: 0.0069, IoU.sand: 0.1407, IoU.sink: 0.2021, IoU.skyscraper: 0.3183, IoU.fireplace: 0.2612, IoU.refrigerator: 0.1892, IoU.grandstand: 0.2116, IoU.path: 0.1079, IoU.stairs: 0.1018, IoU.runway: 0.4893, IoU.case: 0.2450, IoU.pool table: 0.5019, IoU.pillow: 0.1873, IoU.screen door: 0.1706, IoU.stairway: 0.0530, IoU.river: 0.0948, IoU.bridge: 0.0028, IoU.bookcase: 0.0667, IoU.blind: 0.0180, IoU.coffee table: 0.1499, IoU.toilet: 0.2148, IoU.flower: 0.1428, IoU.book: 0.1555, IoU.hill: 0.0346, IoU.bench: 0.0557, IoU.countertop: 0.0753, IoU.stove: 0.1868, IoU.palm: 0.1027, IoU.kitchen island: 0.1596, IoU.computer: 0.0635, IoU.swivel chair: 0.1205, IoU.boat: 0.0543, IoU.bar: 0.0004, IoU.arcade machine: 0.0550, IoU.hovel: 0.0351, IoU.bus: 0.0668, IoU.towel: 0.0235, IoU.light: 0.1098, IoU.truck: 0.0010, IoU.tower: 0.1042, IoU.chandelier: 0.2994, IoU.awning: 0.0035, IoU.streetlight: 0.0028, IoU.booth: 0.0000, IoU.television receiver: 0.1711, IoU.airplane: 0.3153, IoU.dirt track: 0.0000, IoU.apparel: 0.0478, IoU.pole: 0.0196, IoU.land: 0.0000, IoU.bannister: 0.0000, IoU.escalator: 0.0000, IoU.ottoman: 0.0000, IoU.bottle: 0.0000, IoU.buffet: 0.0000, IoU.poster: 0.0000, IoU.stage: 0.0000, IoU.van: 0.0040, IoU.ship: 0.0000, IoU.fountain: 0.0000, IoU.conveyer belt: 0.0000, IoU.canopy: 0.0000, IoU.washer: 0.0028, IoU.plaything: 0.0050, IoU.swimming pool: 0.2438, IoU.stool: 0.0000, IoU.barrel: 0.0000, IoU.basket: 0.0000, IoU.waterfall: 0.4773, IoU.tent: 0.3171, IoU.bag: 0.0000, IoU.minibike: 0.1011, IoU.cradle: 0.2979, IoU.oven: 0.0000, IoU.ball: 0.0017, IoU.food: 0.2743, IoU.step: 0.0000, IoU.tank: 0.0000, IoU.trade name: 0.0335, IoU.microwave: 0.1061, IoU.pot: 0.0000, IoU.animal: 0.0139, IoU.bicycle: 0.0000, IoU.lake: 0.0298, IoU.dishwasher: 0.0275, IoU.screen: 0.2123, IoU.blanket: 0.0000, IoU.sculpture: 0.0000, IoU.hood: 0.0397, IoU.sconce: 0.0198, IoU.vase: 0.0210, IoU.traffic light: 0.0003, IoU.tray: 0.0000, IoU.ashcan: 0.0000, IoU.fan: 0.1251, IoU.pier: 0.0068, IoU.crt screen: 0.0000, IoU.plate: 0.0000, IoU.monitor: 0.0000, IoU.bulletin board: 0.0000, IoU.shower: 0.0000, IoU.radiator: 0.0000, IoU.glass: 0.0000, IoU.clock: 0.0000, IoU.flag: 0.0000, Acc.wall: 0.8183, Acc.building: 0.8928, Acc.sky: 0.9446, Acc.floor: 0.7354, Acc.tree: 0.7873, Acc.ceiling: 0.7305, Acc.road: 0.7987, Acc.bed : 0.7707, Acc.windowpane: 0.6033, Acc.grass: 0.7526, Acc.cabinet: 0.4650, Acc.sidewalk: 0.6320, Acc.person: 0.6408, Acc.earth: 0.3652, Acc.door: 0.0661, Acc.table: 0.3023, Acc.mountain: 0.5138, Acc.plant: 0.4371, Acc.curtain: 0.5299, Acc.chair: 0.3651, Acc.car: 0.7194, Acc.water: 0.3667, Acc.painting: 0.5195, Acc.sofa: 0.4353, Acc.shelf: 0.2372, Acc.house: 0.4256, Acc.sea: 0.6875, Acc.mirror: 0.1420, Acc.rug: 0.2749, Acc.field: 0.4706, Acc.armchair: 0.0084, Acc.seat: 0.3160, Acc.fence: 0.0450, Acc.desk: 0.1550, Acc.rock: 0.2819, Acc.wardrobe: 0.0794, Acc.lamp: 0.1920, Acc.bathtub: 0.1623, Acc.railing: 0.0601, Acc.cushion: 0.2018, Acc.base: 0.0062, Acc.box: 0.0172, Acc.column: 0.0000, Acc.signboard: 0.0479, Acc.chest of drawers: 0.1729, Acc.counter: 0.0070, Acc.sand: 0.2183, Acc.sink: 0.3594, Acc.skyscraper: 0.3765, Acc.fireplace: 0.4307, Acc.refrigerator: 0.2357, Acc.grandstand: 0.3005, Acc.path: 0.1248, Acc.stairs: 0.1139, Acc.runway: 0.7364, Acc.case: 0.4933, Acc.pool table: 0.6341, Acc.pillow: 0.2337, Acc.screen door: 0.1977, Acc.stairway: 0.0701, Acc.river: 0.1644, Acc.bridge: 0.0029, Acc.bookcase: 0.0851, Acc.blind: 0.0184, Acc.coffee table: 0.2761, Acc.toilet: 0.4707, Acc.flower: 0.2360, Acc.book: 0.1832, Acc.hill: 0.0374, Acc.bench: 0.0582, Acc.countertop: 0.0942, Acc.stove: 0.3678, Acc.palm: 0.1179, Acc.kitchen island: 0.1870, Acc.computer: 0.0850, Acc.swivel chair: 0.2080, Acc.boat: 0.0627, Acc.bar: 0.0004, Acc.arcade machine: 0.0623, Acc.hovel: 0.0412, Acc.bus: 0.0719, Acc.towel: 0.0237, Acc.light: 0.1126, Acc.truck: 0.0013, Acc.tower: 0.1066, Acc.chandelier: 0.5031, Acc.awning: 0.0038, Acc.streetlight: 0.0028, Acc.booth: 0.0000, Acc.television receiver: 0.2419, Acc.airplane: 0.3965, Acc.dirt track: 0.0000, Acc.apparel: 0.0566, Acc.pole: 0.0217, Acc.land: 0.0000, Acc.bannister: 0.0000, Acc.escalator: 0.0000, Acc.ottoman: 0.0000, Acc.bottle: 0.0000, Acc.buffet: 0.0000, Acc.poster: 0.0000, Acc.stage: 0.0000, Acc.van: 0.0040, Acc.ship: 0.0000, Acc.fountain: 0.0000, Acc.conveyer belt: 0.0000, Acc.canopy: 0.0000, Acc.washer: 0.0028, Acc.plaything: 0.0054, Acc.swimming pool: 0.4256, Acc.stool: 0.0000, Acc.barrel: 0.0000, Acc.basket: 0.0000, Acc.waterfall: 0.7095, Acc.tent: 0.4717, Acc.bag: 0.0000, Acc.minibike: 0.1168, Acc.cradle: 0.6707, Acc.oven: 0.0000, Acc.ball: 0.0038, Acc.food: 0.3767, Acc.step: 0.0000, Acc.tank: 0.0000, Acc.trade name: 0.0346, Acc.microwave: 0.1233, Acc.pot: 0.0000, Acc.animal: 0.0151, Acc.bicycle: 0.0000, Acc.lake: 0.0354, Acc.dishwasher: 0.0279, Acc.screen: 0.2222, Acc.blanket: 0.0000, Acc.sculpture: 0.0000, Acc.hood: 0.0503, Acc.sconce: 0.0202, Acc.vase: 0.0249, Acc.traffic light: 0.0003, Acc.tray: 0.0000, Acc.ashcan: 0.0000, Acc.fan: 0.1454, Acc.pier: 0.0068, Acc.crt screen: 0.0000, Acc.plate: 0.0000, Acc.monitor: 0.0000, Acc.bulletin board: 0.0000, Acc.shower: 0.0000, Acc.radiator: 0.0000, Acc.glass: 0.0000, Acc.clock: 0.0000, Acc.flag: 0.0000
2021-12-23 16:21:17,443 - mmseg - INFO - Iter [64050/80000]	lr: 2.343e-05, eta: 1:26:26, time: 2.451, data_time: 2.127, memory: 10848, decode.loss_ce: 1.1673, decode.acc_seg: 52.5561, loss: 1.1673
2021-12-23 16:21:34,105 - mmseg - INFO - Iter [64100/80000]	lr: 2.336e-05, eta: 1:26:10, time: 0.333, data_time: 0.007, memory: 10848, decode.loss_ce: 1.1150, decode.acc_seg: 48.8460, loss: 1.1150
2021-12-23 16:21:50,267 - mmseg - INFO - Iter [64150/80000]	lr: 2.330e-05, eta: 1:25:53, time: 0.323, data_time: 0.008, memory: 10848, decode.loss_ce: 1.1292, decode.acc_seg: 48.7623, loss: 1.1292
2021-12-23 16:22:07,223 - mmseg - INFO - Iter [64200/80000]	lr: 2.323e-05, eta: 1:25:37, time: 0.339, data_time: 0.007, memory: 10848, decode.loss_ce: 1.1834, decode.acc_seg: 49.9436, loss: 1.1834
2021-12-23 16:22:23,802 - mmseg - INFO - Iter [64250/80000]	lr: 2.316e-05, eta: 1:25:21, time: 0.331, data_time: 0.008, memory: 10848, decode.loss_ce: 1.1725, decode.acc_seg: 49.8978, loss: 1.1725
2021-12-23 16:22:40,176 - mmseg - INFO - Iter [64300/80000]	lr: 2.310e-05, eta: 1:25:05, time: 0.327, data_time: 0.008, memory: 10848, decode.loss_ce: 1.1656, decode.acc_seg: 50.8126, loss: 1.1656

could you help me to fix this issue
@1104662797
Copy link
Author

the start of log:

2021-12-23 10:33:49,393 - mmseg - INFO - Environment info:
------------------------------------------------------------
sys.platform: linux
Python: 3.9.6 | packaged by conda-forge | (default, Jul 11 2021, 03:39:48) [GCC 9.3.0]
CUDA available: True
GPU 0,1: NVIDIA GeForce RTX 3090
CUDA_HOME: None
GCC: gcc (Ubuntu 7.5.0-6ubuntu2) 7.5.0
PyTorch: 1.8.0
PyTorch compiling details: PyTorch built with:
  - GCC 7.3
  - C++ Version: 201402
  - Intel(R) oneAPI Math Kernel Library Version 2021.3-Product Build 20210617 for Intel(R) 64 architecture applications
  - Intel(R) MKL-DNN v1.7.0 (Git Hash 7aed236906b1f7a05c0917e5257a1af05e9ff683)
  - OpenMP 201511 (a.k.a. OpenMP 4.5)
  - NNPACK is enabled
  - CPU capability usage: AVX2
  - CUDA Runtime 11.1
  - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_37,code=compute_37
  - CuDNN 8.0.5
  - Magma 2.5.2
  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.1, CUDNN_VERSION=8.0.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.8.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, 

TorchVision: 0.2.1
OpenCV: 4.5.3
MMCV: 1.3.18
MMCV Compiler: GCC 7.3
MMCV CUDA Compiler: 11.1
MMSegmentation: 0.19.0+7a1c9a5
------------------------------------------------------------

2021-12-23 10:33:49,394 - mmseg - INFO - Distributed training: True
2021-12-23 10:33:49,597 - mmseg - INFO - Config:
norm_cfg = dict(type='SyncBN', requires_grad=True)
model = dict(
    type='EncoderDecoder',
    pretrained=
    '/home/host/dist1/zx/classification/checkpoints/pvt_tiny/checkpoint.pth',
    backbone=dict(
        type='pvt_tiny',
        depth=50,
        num_stages=4,
        out_indices=(0, 1, 2, 3),
        dilations=(1, 1, 1, 1),
        strides=(1, 2, 2, 2),
        norm_cfg=dict(type='SyncBN', requires_grad=True),
        norm_eval=False,
        style='pytorch',
        contract_dilation=True),
    neck=dict(
        type='FPN',
        in_channels=[64, 128, 320, 512],
        out_channels=256,
        num_outs=4),
    decode_head=dict(
        type='FPNHead',
        in_channels=[256, 256, 256, 256],
        in_index=[0, 1, 2, 3],
        feature_strides=[8, 8, 16, 32],
        channels=128,
        dropout_ratio=0.1,
        num_classes=150,
        norm_cfg=dict(type='SyncBN', requires_grad=True),
        align_corners=False,
        loss_decode=dict(
            type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)),
    train_cfg=dict(),
    test_cfg=dict(mode='whole'))
dataset_type = 'ADE20KDataset'
data_root = 'data/ADEChallengeData2016'
img_norm_cfg = dict(
    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
crop_size = (512, 512)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', reduce_zero_label=True),
    dict(type='Resize', img_scale=(2048, 512), ratio_range=(0.5, 2.0)),
    dict(type='RandomCrop', crop_size=(512, 512), cat_max_ratio=0.75),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PhotoMetricDistortion'),
    dict(
        type='Normalize',
        mean=[123.675, 116.28, 103.53],
        std=[58.395, 57.12, 57.375],
        to_rgb=True),
    dict(type='Pad', size=(512, 512), pad_val=0, seg_pad_val=255),
    dict(type='DefaultFormatBundle'),
    dict(type='Collect', keys=['img', 'gt_semantic_seg'])
]
test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='MultiScaleFlipAug',
        img_scale=(2048, 512),
        flip=False,
        transforms=[
            dict(type='AlignResize', keep_ratio=True, size_divisor=32),
            dict(type='RandomFlip'),
            dict(
                type='Normalize',
                mean=[123.675, 116.28, 103.53],
                std=[58.395, 57.12, 57.375],
                to_rgb=True),
            dict(type='ImageToTensor', keys=['img']),
            dict(type='Collect', keys=['img'])
        ])
]
data = dict(
    samples_per_gpu=4,
    workers_per_gpu=4,
    train=dict(
        type='RepeatDataset',
        times=50,
        dataset=dict(
            type='ADE20KDataset',
            data_root='data/ADEChallengeData2016',
            img_dir='images/training',
            ann_dir='annotations/training',
            pipeline=[
                dict(type='LoadImageFromFile'),
                dict(type='LoadAnnotations', reduce_zero_label=True),
                dict(
                    type='Resize',
                    img_scale=(2048, 512),
                    ratio_range=(0.5, 2.0)),
                dict(
                    type='RandomCrop',
                    crop_size=(512, 512),
                    cat_max_ratio=0.75),
                dict(type='RandomFlip', prob=0.5),
                dict(type='PhotoMetricDistortion'),
                dict(
                    type='Normalize',
                    mean=[123.675, 116.28, 103.53],
                    std=[58.395, 57.12, 57.375],
                    to_rgb=True),
                dict(type='Pad', size=(512, 512), pad_val=0, seg_pad_val=255),
                dict(type='DefaultFormatBundle'),
                dict(type='Collect', keys=['img', 'gt_semantic_seg'])
            ])),
    val=dict(
        type='ADE20KDataset',
        data_root='data/ADEChallengeData2016',
        img_dir='images/validation',
        ann_dir='annotations/validation',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                type='MultiScaleFlipAug',
                img_scale=(2048, 512),
                flip=False,
                transforms=[
                    dict(type='AlignResize', keep_ratio=True, size_divisor=32),
                    dict(type='RandomFlip'),
                    dict(
                        type='Normalize',
                        mean=[123.675, 116.28, 103.53],
                        std=[58.395, 57.12, 57.375],
                        to_rgb=True),
                    dict(type='ImageToTensor', keys=['img']),
                    dict(type='Collect', keys=['img'])
                ])
        ]),
    test=dict(
        type='ADE20KDataset',
        data_root='data/ADEChallengeData2016',
        img_dir='images/validation',
        ann_dir='annotations/validation',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                type='MultiScaleFlipAug',
                img_scale=(2048, 512),
                flip=False,
                transforms=[
                    dict(type='AlignResize', keep_ratio=True, size_divisor=32),
                    dict(type='RandomFlip'),
                    dict(
                        type='Normalize',
                        mean=[123.675, 116.28, 103.53],
                        std=[58.395, 57.12, 57.375],
                        to_rgb=True),
                    dict(type='ImageToTensor', keys=['img']),
                    dict(type='Collect', keys=['img'])
                ])
        ]))
log_config = dict(
    interval=50, hooks=[dict(type='TextLoggerHook', by_epoch=False)])
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = None
resume_from = None
workflow = [('train', 1)]
cudnn_benchmark = True
gpu_multiples = 1
optimizer = dict(type='AdamW', lr=0.0001, weight_decay=0.0001)
optimizer_config = dict()
lr_config = dict(policy='poly', power=0.9, min_lr=0.0, by_epoch=False)
runner = dict(type='IterBasedRunner', max_iters=80000)
checkpoint_config = dict(by_epoch=False, interval=8000)
evaluation = dict(interval=8000, metric='mIoU')
work_dir = './work_dirs/fpn_pvt_t_ade20k_40k'
gpu_ids = range(0, 1)

2021-12-23 10:33:49,947 - mmseg - INFO - EncoderDecoder(
  (backbone): pvt_tiny(
    (patch_embed1): PatchEmbed(
      (proj): Conv2d(3, 64, kernel_size=(4, 4), stride=(4, 4))
      (norm): LayerNorm((64,), eps=1e-05, elementwise_affine=True)
    )
    (pos_drop1): Dropout(p=0.0, inplace=False)
    (block1): ModuleList(
      (0): Block(
        (norm1): LayerNorm((64,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=64, out_features=64, bias=True)
          (kv): Linear(in_features=64, out_features=128, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=64, out_features=64, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
          (sr): Conv2d(64, 64, kernel_size=(8, 8), stride=(8, 8))
          (norm): LayerNorm((64,), eps=1e-05, elementwise_affine=True)
        )
        (drop_path): Identity()
        (norm2): LayerNorm((64,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=64, out_features=512, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=512, out_features=64, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
      (1): Block(
        (norm1): LayerNorm((64,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=64, out_features=64, bias=True)
          (kv): Linear(in_features=64, out_features=128, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=64, out_features=64, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
          (sr): Conv2d(64, 64, kernel_size=(8, 8), stride=(8, 8))
          (norm): LayerNorm((64,), eps=1e-05, elementwise_affine=True)
        )
        (drop_path): DropPath()
        (norm2): LayerNorm((64,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=64, out_features=512, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=512, out_features=64, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
    )
    (patch_embed2): PatchEmbed(
      (proj): Conv2d(64, 128, kernel_size=(2, 2), stride=(2, 2))
      (norm): LayerNorm((128,), eps=1e-05, elementwise_affine=True)
    )
    (pos_drop2): Dropout(p=0.0, inplace=False)
    (block2): ModuleList(
      (0): Block(
        (norm1): LayerNorm((128,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=128, out_features=128, bias=True)
          (kv): Linear(in_features=128, out_features=256, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=128, out_features=128, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
          (sr): Conv2d(128, 128, kernel_size=(4, 4), stride=(4, 4))
          (norm): LayerNorm((128,), eps=1e-05, elementwise_affine=True)
        )
        (drop_path): DropPath()
        (norm2): LayerNorm((128,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=128, out_features=1024, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=1024, out_features=128, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
      (1): Block(
        (norm1): LayerNorm((128,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=128, out_features=128, bias=True)
          (kv): Linear(in_features=128, out_features=256, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=128, out_features=128, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
          (sr): Conv2d(128, 128, kernel_size=(4, 4), stride=(4, 4))
          (norm): LayerNorm((128,), eps=1e-05, elementwise_affine=True)
        )
        (drop_path): DropPath()
        (norm2): LayerNorm((128,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=128, out_features=1024, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=1024, out_features=128, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
    )
    (patch_embed3): PatchEmbed(
      (proj): Conv2d(128, 320, kernel_size=(2, 2), stride=(2, 2))
      (norm): LayerNorm((320,), eps=1e-05, elementwise_affine=True)
    )
    (pos_drop3): Dropout(p=0.0, inplace=False)
    (block3): ModuleList(
      (0): Block(
        (norm1): LayerNorm((320,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=320, out_features=320, bias=True)
          (kv): Linear(in_features=320, out_features=640, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=320, out_features=320, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
          (sr): Conv2d(320, 320, kernel_size=(2, 2), stride=(2, 2))
          (norm): LayerNorm((320,), eps=1e-05, elementwise_affine=True)
        )
        (drop_path): DropPath()
        (norm2): LayerNorm((320,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=320, out_features=1280, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=1280, out_features=320, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
      (1): Block(
        (norm1): LayerNorm((320,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=320, out_features=320, bias=True)
          (kv): Linear(in_features=320, out_features=640, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=320, out_features=320, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
          (sr): Conv2d(320, 320, kernel_size=(2, 2), stride=(2, 2))
          (norm): LayerNorm((320,), eps=1e-05, elementwise_affine=True)
        )
        (drop_path): DropPath()
        (norm2): LayerNorm((320,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=320, out_features=1280, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=1280, out_features=320, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
    )
    (patch_embed4): PatchEmbed(
      (proj): Conv2d(320, 512, kernel_size=(2, 2), stride=(2, 2))
      (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
    )
    (pos_drop4): Dropout(p=0.0, inplace=False)
    (block4): ModuleList(
      (0): Block(
        (norm1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=512, out_features=512, bias=True)
          (kv): Linear(in_features=512, out_features=1024, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=512, out_features=512, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
        )
        (drop_path): DropPath()
        (norm2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=512, out_features=2048, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=2048, out_features=512, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
      (1): Block(
        (norm1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)
        (attn): Attention(
          (q): Linear(in_features=512, out_features=512, bias=True)
          (kv): Linear(in_features=512, out_features=1024, bias=True)
          (attn_drop): Dropout(p=0.0, inplace=False)
          (proj): Linear(in_features=512, out_features=512, bias=True)
          (proj_drop): Dropout(p=0.0, inplace=False)
        )
        (drop_path): DropPath()
        (norm2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)
        (mlp): Mlp(
          (fc1): Linear(in_features=512, out_features=2048, bias=True)
          (act): GELU()
          (fc2): Linear(in_features=2048, out_features=512, bias=True)
          (drop): Dropout(p=0.0, inplace=False)
        )
      )
    )
  )
  (neck): FPN(
    (lateral_convs): ModuleList(
      (0): ConvModule(
        (conv): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1))
      )
      (1): ConvModule(
        (conv): Conv2d(128, 256, kernel_size=(1, 1), stride=(1, 1))
      )
      (2): ConvModule(
        (conv): Conv2d(320, 256, kernel_size=(1, 1), stride=(1, 1))
      )
      (3): ConvModule(
        (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1))
      )
    )
    (fpn_convs): ModuleList(
      (0): ConvModule(
        (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
      )
      (1): ConvModule(
        (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
      )
      (2): ConvModule(
        (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
      )
      (3): ConvModule(
        (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
      )
    )
  )
  init_cfg={'type': 'Xavier', 'layer': 'Conv2d', 'distribution': 'uniform'}
  (decode_head): FPNHead(
    input_transform=multiple_select, ignore_index=255, align_corners=False
    (loss_decode): CrossEntropyLoss()
    (conv_seg): Conv2d(128, 150, kernel_size=(1, 1), stride=(1, 1))
    (dropout): Dropout2d(p=0.1, inplace=False)
    (scale_heads): ModuleList(
      (0): Sequential(
        (0): ConvModule(
          (conv): Conv2d(256, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (activate): ReLU(inplace=True)
        )
      )
      (1): Sequential(
        (0): ConvModule(
          (conv): Conv2d(256, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (activate): ReLU(inplace=True)
        )
      )
      (2): Sequential(
        (0): ConvModule(
          (conv): Conv2d(256, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (activate): ReLU(inplace=True)
        )
        (1): Upsample()
      )
      (3): Sequential(
        (0): ConvModule(
          (conv): Conv2d(256, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (activate): ReLU(inplace=True)
        )
        (1): Upsample()
        (2): ConvModule(
          (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn): SyncBatchNorm(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (activate): ReLU(inplace=True)
        )
        (3): Upsample()
      )
    )
  )
  init_cfg={'type': 'Normal', 'std': 0.01, 'override': {'name': 'conv_seg'}}
)
2021-12-23 10:33:50,289 - mmseg - INFO - Loaded 20210 images
2021-12-23 10:33:55,045 - mmseg - INFO - Loaded 2000 images
2021-12-23 10:33:55,046 - mmseg - INFO - Start running, host: host@ubuntu20, work_dir: /home/host/mounted1/zxDet/mmsegmentation/work_dirs/fpn_pvt_t_ade20k_40k
2021-12-23 10:33:55,046 - mmseg - INFO - Hooks will be executed in the following order:
before_run:
(VERY_HIGH   ) PolyLrUpdaterHook                  
(NORMAL      ) CheckpointHook                     
(LOW         ) DistEvalHook                       
(VERY_LOW    ) TextLoggerHook                     
 -------------------- 
before_train_epoch:
(VERY_HIGH   ) PolyLrUpdaterHook                  
(LOW         ) IterTimerHook                      
(LOW         ) DistEvalHook                       
(VERY_LOW    ) TextLoggerHook                     
 -------------------- 
before_train_iter:
(VERY_HIGH   ) PolyLrUpdaterHook                  
(LOW         ) IterTimerHook                      
(LOW         ) DistEvalHook                       
 -------------------- 
after_train_iter:
(ABOVE_NORMAL) OptimizerHook                      
(NORMAL      ) CheckpointHook                     
(LOW         ) IterTimerHook                      
(LOW         ) DistEvalHook                       
(VERY_LOW    ) TextLoggerHook                     
 -------------------- 
after_train_epoch:
(NORMAL      ) CheckpointHook                     
(LOW         ) DistEvalHook                       
(VERY_LOW    ) TextLoggerHook                     
 -------------------- 
before_val_epoch:
(LOW         ) IterTimerHook                      
(VERY_LOW    ) TextLoggerHook                     
 -------------------- 
before_val_iter:
(LOW         ) IterTimerHook                      
 -------------------- 
after_val_iter:
(LOW         ) IterTimerHook                      
 -------------------- 
after_val_epoch:
(VERY_LOW    ) TextLoggerHook                     
 -------------------- 
after_run:
(VERY_LOW    ) TextLoggerHook                     
 -------------------- 
2021-12-23 10:33:55,046 - mmseg - INFO - workflow: [('train', 1)], max: 80000 iters

@czczup
Copy link
Collaborator

czczup commented Jan 7, 2022

The pretrained weight seems not to be loaded because I didn't see a warning about the head.weight.

@Leisureon
Copy link

I use pvt_tiny for this two jobs. But the mIoU is very low. we train with this script on two rtx3090, I just change the gpu_multiples to 1:

_base_ = [
    '../../_base_/models/fpn_r50.py',
    '../../_base_/datasets/ade20k.py',
    '../../_base_/default_runtime.py'
]
# model settings
model = dict(
    type='EncoderDecoder',
    # pretrained='pretrained/pvt_tiny.pth',
    pretrained='/home/host/dist1/zx/classification/checkpoints/pvt_tiny/checkpoint.pth',
    backbone=dict(
        type='pvt_tiny',
        style='pytorch'),
    neck=dict(in_channels=[64, 128, 320, 512]),
    decode_head=dict(num_classes=150))


gpu_multiples = 1  # we use 8 gpu instead of 4 in mmsegmentation, so lr*2 and max_iters/2
# optimizeiiiiiir
optimizer = dict(type='AdamW', lr=0.0001*gpu_multiples, weight_decay=0.0001)
optimizer_config = dict()
# learning policy
lr_config = dict(policy='poly', power=0.9, min_lr=0.0, by_epoch=False)
# runtime settings
runner = dict(type='IterBasedRunner', max_iters=80000//gpu_multiples)
checkpoint_config = dict(by_epoch=False, interval=8000//gpu_multiples)
evaluation = dict(interval=8000//gpu_multiples, metric='mIoU')

the part of log I get :

       Class        |  IoU  |  Acc  |
+---------------------+-------+-------+
|         wall        |  54.7 | 81.83 |
|       building      | 67.78 | 89.28 |
|         sky         |  89.7 | 94.46 |
|        floor        | 57.27 | 73.54 |
|         tree        | 58.49 | 78.73 |
|       ceiling       | 63.11 | 73.05 |
|         road        | 64.98 | 79.87 |
|         bed         | 48.05 | 77.07 |
|      windowpane     | 40.18 | 60.33 |
|        grass        | 57.12 | 75.26 |
|       cabinet       | 32.02 |  46.5 |
|       sidewalk      | 41.35 |  63.2 |
|        person       | 40.43 | 64.08 |
|        earth        | 27.29 | 36.52 |
|         door        |  5.95 |  6.61 |
|        table        | 18.18 | 30.23 |
|       mountain      | 31.44 | 51.38 |
|        plant        | 33.12 | 43.71 |
|       curtain       | 38.04 | 52.99 |
|        chair        | 21.51 | 36.51 |
|         car         |  51.0 | 71.94 |
|        water        | 29.28 | 36.67 |
|       painting      | 31.18 | 51.95 |
|         sofa        | 24.42 | 43.53 |
|        shelf        | 13.72 | 23.72 |
|        house        | 28.35 | 42.56 |
|         sea         | 42.37 | 68.75 |
|        mirror       | 11.43 |  14.2 |
|         rug         | 23.17 | 27.49 |
|        field        | 22.71 | 47.06 |
|       armchair      |  0.78 |  0.84 |
|         seat        | 25.22 |  31.6 |
|        fence        |  4.03 |  4.5  |
|         desk        | 10.89 |  15.5 |
|         rock        |  16.8 | 28.19 |
|       wardrobe      |  7.48 |  7.94 |
|         lamp        | 14.81 |  19.2 |
|       bathtub       | 13.31 | 16.23 |
|       railing       |  5.92 |  6.01 |
|       cushion       | 13.75 | 20.18 |
|         base        |  0.55 |  0.62 |
|         box         |  1.56 |  1.72 |
|        column       |  0.0  |  0.0  |
|      signboard      |  4.54 |  4.79 |
|   chest of drawers  | 12.02 | 17.29 |
|       counter       |  0.69 |  0.7  |
|         sand        | 14.07 | 21.83 |
|         sink        | 20.21 | 35.94 |
|      skyscraper     | 31.83 | 37.65 |
|      fireplace      | 26.12 | 43.07 |
|     refrigerator    | 18.92 | 23.57 |
|      grandstand     | 21.16 | 30.05 |
|         path        | 10.79 | 12.48 |
|        stairs       | 10.18 | 11.39 |
|        runway       | 48.93 | 73.64 |
|         case        |  24.5 | 49.33 |
|      pool table     | 50.19 | 63.41 |
|        pillow       | 18.73 | 23.37 |
|     screen door     | 17.06 | 19.77 |
|       stairway      |  5.3  |  7.01 |
|        river        |  9.48 | 16.44 |
|        bridge       |  0.28 |  0.29 |
|       bookcase      |  6.67 |  8.51 |
|        blind        |  1.8  |  1.84 |
|     coffee table    | 14.99 | 27.61 |
|        toilet       | 21.48 | 47.07 |
|        flower       | 14.28 |  23.6 |
|         book        | 15.55 | 18.32 |
|         hill        |  3.46 |  3.74 |
|        bench        |  5.57 |  5.82 |
|      countertop     |  7.53 |  9.42 |
|        stove        | 18.68 | 36.78 |
|         palm        | 10.27 | 11.79 |
|    kitchen island   | 15.96 |  18.7 |
|       computer      |  6.35 |  8.5  |
|     swivel chair    | 12.05 |  20.8 |
|         boat        |  5.43 |  6.27 |
|         bar         |  0.04 |  0.04 |
|    arcade machine   |  5.5  |  6.23 |
|        hovel        |  3.51 |  4.12 |
|         bus         |  6.68 |  7.19 |
|        towel        |  2.35 |  2.37 |
|        light        | 10.98 | 11.26 |
|        truck        |  0.1  |  0.13 |
|        tower        | 10.42 | 10.66 |
|      chandelier     | 29.94 | 50.31 |
|        awning       |  0.35 |  0.38 |
|     streetlight     |  0.28 |  0.28 |
|        booth        |  0.0  |  0.0  |
| television receiver | 17.11 | 24.19 |
|       airplane      | 31.53 | 39.65 |
|      dirt track     |  0.0  |  0.0  |
|       apparel       |  4.78 |  5.66 |
|         pole        |  1.96 |  2.17 |
|         land        |  0.0  |  0.0  |
|      bannister      |  0.0  |  0.0  |
|      escalator      |  0.0  |  0.0  |
|       ottoman       |  0.0  |  0.0  |
|        bottle       |  0.0  |  0.0  |
|        buffet       |  0.0  |  0.0  |
|        poster       |  0.0  |  0.0  |
|        stage        |  0.0  |  0.0  |
|         van         |  0.4  |  0.4  |
|         ship        |  0.0  |  0.0  |
|       fountain      |  0.0  |  0.0  |
|    conveyer belt    |  0.0  |  0.0  |
|        canopy       |  0.0  |  0.0  |
|        washer       |  0.28 |  0.28 |
|      plaything      |  0.5  |  0.54 |
|    swimming pool    | 24.38 | 42.56 |
|        stool        |  0.0  |  0.0  |
|        barrel       |  0.0  |  0.0  |
|        basket       |  0.0  |  0.0  |
|      waterfall      | 47.73 | 70.95 |
|         tent        | 31.71 | 47.17 |
|         bag         |  0.0  |  0.0  |
|       minibike      | 10.11 | 11.68 |
|        cradle       | 29.79 | 67.07 |
|         oven        |  0.0  |  0.0  |
|         ball        |  0.17 |  0.38 |
|         food        | 27.43 | 37.67 |
|         step        |  0.0  |  0.0  |
|         tank        |  0.0  |  0.0  |
|      trade name     |  3.35 |  3.46 |
|      microwave      | 10.61 | 12.33 |
|         pot         |  0.0  |  0.0  |
|        animal       |  1.39 |  1.51 |
|       bicycle       |  0.0  |  0.0  |
|         lake        |  2.98 |  3.54 |
|      dishwasher     |  2.75 |  2.79 |
|        screen       | 21.23 | 22.22 |
|       blanket       |  0.0  |  0.0  |
|      sculpture      |  0.0  |  0.0  |
|         hood        |  3.97 |  5.03 |
|        sconce       |  1.98 |  2.02 |
|         vase        |  2.1  |  2.49 |
|    traffic light    |  0.03 |  0.03 |
|         tray        |  0.0  |  0.0  |
|        ashcan       |  0.0  |  0.0  |
|         fan         | 12.51 | 14.54 |
|         pier        |  0.68 |  0.68 |
|      crt screen     |  0.0  |  0.0  |
|        plate        |  0.0  |  0.0  |
|       monitor       |  0.0  |  0.0  |
|    bulletin board   |  0.0  |  0.0  |
|        shower       |  0.0  |  0.0  |
|       radiator      |  0.0  |  0.0  |
|        glass        |  0.0  |  0.0  |
|        clock        |  0.0  |  0.0  |
|         flag        |  0.0  |  0.0  |
+---------------------+-------+-------+
2021-12-23 16:21:00,788 - mmseg - INFO - Summary:
2021-12-23 16:21:00,789 - mmseg - INFO - 
+-------+-------+-------+
|  aAcc |  mIoU |  mAcc |
+-------+-------+-------+
| 65.06 | 14.33 | 20.34 |
+-------+-------+-------+
2021-12-23 16:21:00,804 - mmseg - INFO - Exp name: fpn_pvt_t_ade20k_40k.py
2021-12-23 16:21:00,805 - mmseg - INFO - Iter(val) [1000]	aAcc: 0.6506, mIoU: 0.1433, mAcc: 0.2034, IoU.wall: 0.5470, IoU.building: 0.6778, IoU.sky: 0.8970, IoU.floor: 0.5727, IoU.tree: 0.5849, IoU.ceiling: 0.6311, IoU.road: 0.6498, IoU.bed : 0.4805, IoU.windowpane: 0.4018, IoU.grass: 0.5712, IoU.cabinet: 0.3202, IoU.sidewalk: 0.4135, IoU.person: 0.4043, IoU.earth: 0.2729, IoU.door: 0.0595, IoU.table: 0.1818, IoU.mountain: 0.3144, IoU.plant: 0.3312, IoU.curtain: 0.3804, IoU.chair: 0.2151, IoU.car: 0.5100, IoU.water: 0.2928, IoU.painting: 0.3118, IoU.sofa: 0.2442, IoU.shelf: 0.1372, IoU.house: 0.2835, IoU.sea: 0.4237, IoU.mirror: 0.1143, IoU.rug: 0.2317, IoU.field: 0.2271, IoU.armchair: 0.0078, IoU.seat: 0.2522, IoU.fence: 0.0403, IoU.desk: 0.1089, IoU.rock: 0.1680, IoU.wardrobe: 0.0748, IoU.lamp: 0.1481, IoU.bathtub: 0.1331, IoU.railing: 0.0592, IoU.cushion: 0.1375, IoU.base: 0.0055, IoU.box: 0.0156, IoU.column: 0.0000, IoU.signboard: 0.0454, IoU.chest of drawers: 0.1202, IoU.counter: 0.0069, IoU.sand: 0.1407, IoU.sink: 0.2021, IoU.skyscraper: 0.3183, IoU.fireplace: 0.2612, IoU.refrigerator: 0.1892, IoU.grandstand: 0.2116, IoU.path: 0.1079, IoU.stairs: 0.1018, IoU.runway: 0.4893, IoU.case: 0.2450, IoU.pool table: 0.5019, IoU.pillow: 0.1873, IoU.screen door: 0.1706, IoU.stairway: 0.0530, IoU.river: 0.0948, IoU.bridge: 0.0028, IoU.bookcase: 0.0667, IoU.blind: 0.0180, IoU.coffee table: 0.1499, IoU.toilet: 0.2148, IoU.flower: 0.1428, IoU.book: 0.1555, IoU.hill: 0.0346, IoU.bench: 0.0557, IoU.countertop: 0.0753, IoU.stove: 0.1868, IoU.palm: 0.1027, IoU.kitchen island: 0.1596, IoU.computer: 0.0635, IoU.swivel chair: 0.1205, IoU.boat: 0.0543, IoU.bar: 0.0004, IoU.arcade machine: 0.0550, IoU.hovel: 0.0351, IoU.bus: 0.0668, IoU.towel: 0.0235, IoU.light: 0.1098, IoU.truck: 0.0010, IoU.tower: 0.1042, IoU.chandelier: 0.2994, IoU.awning: 0.0035, IoU.streetlight: 0.0028, IoU.booth: 0.0000, IoU.television receiver: 0.1711, IoU.airplane: 0.3153, IoU.dirt track: 0.0000, IoU.apparel: 0.0478, IoU.pole: 0.0196, IoU.land: 0.0000, IoU.bannister: 0.0000, IoU.escalator: 0.0000, IoU.ottoman: 0.0000, IoU.bottle: 0.0000, IoU.buffet: 0.0000, IoU.poster: 0.0000, IoU.stage: 0.0000, IoU.van: 0.0040, IoU.ship: 0.0000, IoU.fountain: 0.0000, IoU.conveyer belt: 0.0000, IoU.canopy: 0.0000, IoU.washer: 0.0028, IoU.plaything: 0.0050, IoU.swimming pool: 0.2438, IoU.stool: 0.0000, IoU.barrel: 0.0000, IoU.basket: 0.0000, IoU.waterfall: 0.4773, IoU.tent: 0.3171, IoU.bag: 0.0000, IoU.minibike: 0.1011, IoU.cradle: 0.2979, IoU.oven: 0.0000, IoU.ball: 0.0017, IoU.food: 0.2743, IoU.step: 0.0000, IoU.tank: 0.0000, IoU.trade name: 0.0335, IoU.microwave: 0.1061, IoU.pot: 0.0000, IoU.animal: 0.0139, IoU.bicycle: 0.0000, IoU.lake: 0.0298, IoU.dishwasher: 0.0275, IoU.screen: 0.2123, IoU.blanket: 0.0000, IoU.sculpture: 0.0000, IoU.hood: 0.0397, IoU.sconce: 0.0198, IoU.vase: 0.0210, IoU.traffic light: 0.0003, IoU.tray: 0.0000, IoU.ashcan: 0.0000, IoU.fan: 0.1251, IoU.pier: 0.0068, IoU.crt screen: 0.0000, IoU.plate: 0.0000, IoU.monitor: 0.0000, IoU.bulletin board: 0.0000, IoU.shower: 0.0000, IoU.radiator: 0.0000, IoU.glass: 0.0000, IoU.clock: 0.0000, IoU.flag: 0.0000, Acc.wall: 0.8183, Acc.building: 0.8928, Acc.sky: 0.9446, Acc.floor: 0.7354, Acc.tree: 0.7873, Acc.ceiling: 0.7305, Acc.road: 0.7987, Acc.bed : 0.7707, Acc.windowpane: 0.6033, Acc.grass: 0.7526, Acc.cabinet: 0.4650, Acc.sidewalk: 0.6320, Acc.person: 0.6408, Acc.earth: 0.3652, Acc.door: 0.0661, Acc.table: 0.3023, Acc.mountain: 0.5138, Acc.plant: 0.4371, Acc.curtain: 0.5299, Acc.chair: 0.3651, Acc.car: 0.7194, Acc.water: 0.3667, Acc.painting: 0.5195, Acc.sofa: 0.4353, Acc.shelf: 0.2372, Acc.house: 0.4256, Acc.sea: 0.6875, Acc.mirror: 0.1420, Acc.rug: 0.2749, Acc.field: 0.4706, Acc.armchair: 0.0084, Acc.seat: 0.3160, Acc.fence: 0.0450, Acc.desk: 0.1550, Acc.rock: 0.2819, Acc.wardrobe: 0.0794, Acc.lamp: 0.1920, Acc.bathtub: 0.1623, Acc.railing: 0.0601, Acc.cushion: 0.2018, Acc.base: 0.0062, Acc.box: 0.0172, Acc.column: 0.0000, Acc.signboard: 0.0479, Acc.chest of drawers: 0.1729, Acc.counter: 0.0070, Acc.sand: 0.2183, Acc.sink: 0.3594, Acc.skyscraper: 0.3765, Acc.fireplace: 0.4307, Acc.refrigerator: 0.2357, Acc.grandstand: 0.3005, Acc.path: 0.1248, Acc.stairs: 0.1139, Acc.runway: 0.7364, Acc.case: 0.4933, Acc.pool table: 0.6341, Acc.pillow: 0.2337, Acc.screen door: 0.1977, Acc.stairway: 0.0701, Acc.river: 0.1644, Acc.bridge: 0.0029, Acc.bookcase: 0.0851, Acc.blind: 0.0184, Acc.coffee table: 0.2761, Acc.toilet: 0.4707, Acc.flower: 0.2360, Acc.book: 0.1832, Acc.hill: 0.0374, Acc.bench: 0.0582, Acc.countertop: 0.0942, Acc.stove: 0.3678, Acc.palm: 0.1179, Acc.kitchen island: 0.1870, Acc.computer: 0.0850, Acc.swivel chair: 0.2080, Acc.boat: 0.0627, Acc.bar: 0.0004, Acc.arcade machine: 0.0623, Acc.hovel: 0.0412, Acc.bus: 0.0719, Acc.towel: 0.0237, Acc.light: 0.1126, Acc.truck: 0.0013, Acc.tower: 0.1066, Acc.chandelier: 0.5031, Acc.awning: 0.0038, Acc.streetlight: 0.0028, Acc.booth: 0.0000, Acc.television receiver: 0.2419, Acc.airplane: 0.3965, Acc.dirt track: 0.0000, Acc.apparel: 0.0566, Acc.pole: 0.0217, Acc.land: 0.0000, Acc.bannister: 0.0000, Acc.escalator: 0.0000, Acc.ottoman: 0.0000, Acc.bottle: 0.0000, Acc.buffet: 0.0000, Acc.poster: 0.0000, Acc.stage: 0.0000, Acc.van: 0.0040, Acc.ship: 0.0000, Acc.fountain: 0.0000, Acc.conveyer belt: 0.0000, Acc.canopy: 0.0000, Acc.washer: 0.0028, Acc.plaything: 0.0054, Acc.swimming pool: 0.4256, Acc.stool: 0.0000, Acc.barrel: 0.0000, Acc.basket: 0.0000, Acc.waterfall: 0.7095, Acc.tent: 0.4717, Acc.bag: 0.0000, Acc.minibike: 0.1168, Acc.cradle: 0.6707, Acc.oven: 0.0000, Acc.ball: 0.0038, Acc.food: 0.3767, Acc.step: 0.0000, Acc.tank: 0.0000, Acc.trade name: 0.0346, Acc.microwave: 0.1233, Acc.pot: 0.0000, Acc.animal: 0.0151, Acc.bicycle: 0.0000, Acc.lake: 0.0354, Acc.dishwasher: 0.0279, Acc.screen: 0.2222, Acc.blanket: 0.0000, Acc.sculpture: 0.0000, Acc.hood: 0.0503, Acc.sconce: 0.0202, Acc.vase: 0.0249, Acc.traffic light: 0.0003, Acc.tray: 0.0000, Acc.ashcan: 0.0000, Acc.fan: 0.1454, Acc.pier: 0.0068, Acc.crt screen: 0.0000, Acc.plate: 0.0000, Acc.monitor: 0.0000, Acc.bulletin board: 0.0000, Acc.shower: 0.0000, Acc.radiator: 0.0000, Acc.glass: 0.0000, Acc.clock: 0.0000, Acc.flag: 0.0000
2021-12-23 16:21:17,443 - mmseg - INFO - Iter [64050/80000]	lr: 2.343e-05, eta: 1:26:26, time: 2.451, data_time: 2.127, memory: 10848, decode.loss_ce: 1.1673, decode.acc_seg: 52.5561, loss: 1.1673
2021-12-23 16:21:34,105 - mmseg - INFO - Iter [64100/80000]	lr: 2.336e-05, eta: 1:26:10, time: 0.333, data_time: 0.007, memory: 10848, decode.loss_ce: 1.1150, decode.acc_seg: 48.8460, loss: 1.1150
2021-12-23 16:21:50,267 - mmseg - INFO - Iter [64150/80000]	lr: 2.330e-05, eta: 1:25:53, time: 0.323, data_time: 0.008, memory: 10848, decode.loss_ce: 1.1292, decode.acc_seg: 48.7623, loss: 1.1292
2021-12-23 16:22:07,223 - mmseg - INFO - Iter [64200/80000]	lr: 2.323e-05, eta: 1:25:37, time: 0.339, data_time: 0.007, memory: 10848, decode.loss_ce: 1.1834, decode.acc_seg: 49.9436, loss: 1.1834
2021-12-23 16:22:23,802 - mmseg - INFO - Iter [64250/80000]	lr: 2.316e-05, eta: 1:25:21, time: 0.331, data_time: 0.008, memory: 10848, decode.loss_ce: 1.1725, decode.acc_seg: 49.8978, loss: 1.1725
2021-12-23 16:22:40,176 - mmseg - INFO - Iter [64300/80000]	lr: 2.310e-05, eta: 1:25:05, time: 0.327, data_time: 0.008, memory: 10848, decode.loss_ce: 1.1656, decode.acc_seg: 50.8126, loss: 1.1656

could you help me to fix this issue

Where can I download the pre training weight？Thank you！

@shy-star
Copy link

The pretrained weight seems not to be loaded because I didn't see a warning about the head.weight.

I would like to know how the miou was obtained in the paper, I used the method in the paper and the trained miou results were around 15.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
4 participants
@czczup @shy-star @1104662797 @Leisureon