Include each class metrics in logs

[sshuair]

Add each class metric to the .log and .log.json.

Each class metric format is metric.class:0.43, such as IoU.Persion: 0.2432, Acc:Persion: 0.5434.

This PR Fix the iussue #397

For now the output should be like this:

xxx.log

2021-03-27 23:50:31,241 - mmseg - INFO - Iter(val) [10] mIoU: 0.0492, mAcc: 0.0714, aAcc: 0.6881, IoU.background: 0.6881, IoU.aeroplane: 0.0000, IoU.bicycle: 0.0000, IoU.bird: nan, IoU.boat: 0.0000, IoU.bottle: 0.0000, IoU.bus: nan, IoU.car: 0.0000, IoU.cat: nan, IoU.chair: 0.0000, IoU.cow: 0.0000, IoU.diningtable: nan, IoU.dog: nan, IoU.horse: 0.0000, IoU.motorbike: nan, IoU.person: 0.0000, IoU.pottedplant: nan, IoU.sheep: 0.0000, IoU.sofa: 0.0000, IoU.train: 0.0000, IoU.tvmonitor: 0.0000, Acc.background: 1.0000, Acc.aeroplane: 0.0000, Acc.bicycle: 0.0000, Acc.bird: nan, Acc.boat: 0.0000, Acc.bottle: 0.0000, Acc.bus: nan, Acc.car: 0.0000, Acc.cat: nan, Acc.chair: 0.0000, Acc.cow: 0.0000, Acc.diningtable: nan, Acc.dog: nan, Acc.horse: 0.0000, Acc.motorbike: nan, Acc.person: 0.0000, Acc.pottedplant: nan, Acc.sheep: 0.0000, Acc.sofa: 0.0000, Acc.train: 0.0000, Acc.tvmonitor: 0.0000

| person      | 0.0   | 0.0   |
| pottedplant | nan   | nan   |
| sheep       | 0.0   | 0.0   |
| sofa        | 0.0   | 0.0   |
| train       | 0.0   | 0.0   |
| tvmonitor   | 0.0   | 0.0   |
+-------------+-------+-------+
2021-03-27 23:50:31,240 - mmseg - INFO - Summary:
2021-03-27 23:50:31,240 - mmseg - INFO -
+--------+------+------+-------+
| Scope  | mIoU | mAcc | aAcc  |
+--------+------+------+-------+
| global | 4.92 | 7.14 | 68.81 |
+--------+------+------+-------+
2021-03-27 23:50:31,241 - mmseg - INFO - Iter(val) [10] mIoU: 0.0492, mAcc: 0.0714, aAcc: 0.6881, IoU.background: 0.6881, IoU.aeroplane: 0.0000, IoU.bicycle: 0.0000, IoU.bird: nan, IoU.boat: 0.0000, IoU.bottle: 0.0000, IoU.bus: nan, IoU.car: 0.0000, IoU.cat: nan, IoU.chair: 0.0000, IoU.cow: 0.0000, IoU.diningtable: nan, IoU.dog: nan, IoU.horse: 0.0000, IoU.motorbike: nan, IoU.person: 0.0000, IoU.pottedplant: nan, IoU.sheep: 0.0000, IoU.sofa: 0.0000, IoU.train: 0.0000, IoU.tvmonitor: 0.0000, Acc.background: 1.0000, Acc.aeroplane: 0.0000, Acc.bicycle: 0.0000, Acc.bird: nan, Acc.boat: 0.0000, Acc.bottle: 0.0000, Acc.bus: nan, Acc.car: 0.0000, Acc.cat: nan, Acc.chair: 0.0000, Acc.cow: 0.0000, Acc.diningtable: nan, Acc.dog: nan, Acc.horse: 0.0000, Acc.motorbike: nan, Acc.person: 0.0000, Acc.pottedplant: nan, Acc.sheep: 0.0000, Acc.sofa: 0.0000, Acc.train: 0.0000, Acc.tvmonitor: 0.0000
2021-03-27 23:50:31,745 - mmseg - INFO - Iter [11/40000]        lr: 9.998e-03, eta: 12:59:57, time: 6.345, data_time: 5.846, memory: 3569, decode.loss_seg: 1.6204, decode.acc_seg: 63.9891, loss: 1.6204
2021-03-27 23:50:32,246 - mmseg - INFO - Iter [12/40000]        lr: 9.998e-03, eta: 12:22:44, time: 0.501, data_time: 0.005, memory: 3569, decode.loss_seg: 2.7463, decode.acc_seg: 46.6898, loss: 2.7463
2021-03-27 23:50:32,747 - mmseg - INFO - Iter [13/40000]        lr: 9.997e-03, eta: 11:51:16, time: 0.501, data_time: 0.005, memory: 3569, decode.loss_seg: 1.5349, decode.acc_seg: 69.0987, loss: 1.5349
2021-03-27 23:50:33,249 - mmseg - INFO - Iter [14/40000]        lr: 9.997e-03, eta: 11:24:20, time: 0.502, data_time: 0.005, memory: 3569, decode.loss_seg: 2.2271, decode.acc_seg: 39.2558, loss: 2.2271
2021-03-27 23:50:33,752 - mmseg - INFO - Iter [15/40000]        lr: 9.997e-03, eta: 11:01:03, time: 0.503, data_time: 0.005, memory: 3569, decode.loss_seg: 0.6409, decode.acc_seg: 81.1710, loss: 0.6409
2021-03-27 23:50:34,257 - mmseg - INFO - Iter [16/40000]        lr: 9.997e-03, eta: 10:40:45, time: 0.505, data_time: 0.005, memory: 3569, decode.loss_seg: 1.1608, decode.acc_seg: 62.2077, loss: 1.1608

xxx.log.json

...
{"mode": "val", "epoch": 1, "iter": 10, "lr": 0.01, "mIoU": 0.0577, "mAcc": 0.0977, "aAcc": 0.5454, "IoU.road": 0.7423, "IoU.sidewalk": 0.0, "IoU.building": 0.3544, "IoU.wall": 0.0, "IoU.fence": 0.0, "IoU.pole": 0.0, "IoU.traffic light": 0.0, "IoU.traffic sign": 0.0, "IoU.vegetation": 0.0, "IoU.terrain": 0.0, "IoU.sky": 0.0, "IoU.person": 0.0, "IoU.rider": 0.0, "IoU.car": 0.0, "IoU.truck": 0.0, "IoU.bus": 0.0, "IoU.train": 0.0, "IoU.motorcycle": 0.0, "IoU.bicycle": 0.0, "Acc.road": 0.8795, "Acc.sidewalk": 0.0, "Acc.building": 0.977, "Acc.wall": 0.0, "Acc.fence": 0.0, "Acc.pole": 0.0, "Acc.traffic light": 0.0, "Acc.traffic sign": 0.0, "Acc.vegetation": 0.0, "Acc.terrain": 0.0, "Acc.sky": 0.0, "Acc.person": 0.0, "Acc.rider": 0.0, "Acc.car": 0.0, "Acc.truck": 0.0, "Acc.bus": 0.0, "Acc.train": 0.0, "Acc.motorcycle": 0.0, "Acc.bicycle": 0.0}
...

{"mode": "train", "epoch": 1, "iter": 7, "lr": 0.01, "memory": 3569, "data_time": 0.00477, "decode.loss_seg": 1.78463, "decode.acc_seg": 51.22833, "loss": 1.78463, "time": 0.50375}
{"mode": "train", "epoch": 1, "iter": 8, "lr": 0.01, "memory": 3569, "data_time": 0.00475, "decode.loss_seg": 1.70086, "decode.acc_seg": 60.25438, "loss": 1.70086, "time": 0.50268}
{"mode": "train", "epoch": 1, "iter": 9, "lr": 0.01, "memory": 3569, "data_time": 0.00477, "decode.loss_seg": 1.27715, "decode.acc_seg": 56.3323, "loss": 1.27715, "time": 0.50368}
{"mode": "val", "epoch": 1, "iter": 10, "lr": 0.01, "mIoU": 0.0492, "mAcc": 0.0714, "aAcc": 0.6881, "IoU.background": 0.6881, "IoU.aeroplane": 0.0, "IoU.bicycle": 0.0, "IoU.bird": NaN, "IoU.boat": 0.0, "IoU.bottle": 0.0, "IoU.bus": NaN, "IoU.car": 0.0, "IoU.cat": NaN, "IoU.chair": 0.0, "IoU.cow": 0.0, "IoU.diningtable": NaN, "IoU.dog": NaN, "IoU.horse": 0.0, "IoU.motorbike": NaN, "IoU.person": 0.0, "IoU.pottedplant": NaN, "IoU.sheep": 0.0, "IoU.sofa": 0.0, "IoU.train": 0.0, "IoU.tvmonitor": 0.0, "Acc.background": 1.0, "Acc.aeroplane": 0.0, "Acc.bicycle": 0.0, "Acc.bird": NaN, "Acc.boat": 0.0, "Acc.bottle": 0.0, "Acc.bus": NaN, "Acc.car": 0.0, "Acc.cat": NaN, "Acc.chair": 0.0, "Acc.cow": 0.0, "Acc.diningtable": NaN, "Acc.dog": NaN, "Acc.horse": 0.0, "Acc.motorbike": NaN, "Acc.person": 0.0, "Acc.pottedplant": NaN, "Acc.sheep": 0.0, "Acc.sofa": 0.0, "Acc.train": 0.0, "Acc.tvmonitor": 0.0}
{"mode": "train", "epoch": 1, "iter": 11, "lr": 0.01, "memory": 3569, "data_time": 5.84578, "decode.loss_seg": 1.62035, "decode.acc_seg": 63.98907, "loss": 1.62035, "time": 6.34452}

[xvjiarui]

Hi @sshuair
Thanks very much for your contribution. It seems that unittest has failed. Could you please fix it?

[codecov]

Codecov Report

Merging #445 (b8985d8) into master (6acd77e) will decrease coverage by 0.05%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #445      +/-   ##
==========================================
- Coverage   86.43%   86.37%   -0.06%     
==========================================
  Files          97       97              
  Lines        4960     4963       +3     
  Branches      805      807       +2     
==========================================
  Hits         4287     4287              
- Misses        518      519       +1     
- Partials      155      157       +2     

Flag	Coverage Δ
unittests	86.37% <100.00%> (-0.06%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
mmseg/datasets/custom.py	88.81% <100.00%> (+0.21%)	⬆️
mmseg/datasets/pipelines/transforms.py	97.12% <0.00%> (-0.87%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6acd77e...b8985d8. Read the comment docs.

[sshuair]

@xvjiarui Hi the unittest has passed. But the codecov still has some problem. I don't know how to fix it.

[xvjiarui]

Hi @sshuair
It seems that this part it not tested.

[sshuair]

@xvjiarui which part? It seems that the transforms.py caused the problem. But I didn’t modify this file.