Update pretrained models of STARK on LaSOT and TrackingNet

[JingweiZhang12]

Motivation

The previous models of stark on LaSOT and TrackingNet are trained on part of LaSOT dataset and the wrong split of GOT10k. This time, we train the model on the right settings and update the models and docs.
BTW, we update the right memory occupation of training STARK.

[codecov]

Codecov Report

Merging #503 (79c66bc) into master (503756c) will increase coverage by 0.50%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #503      +/-   ##
==========================================
+ Coverage   73.15%   73.65%   +0.50%     
==========================================
  Files         126      126              
  Lines        7361     7361              
  Branches     1379     1379              
==========================================
+ Hits         5385     5422      +37     
+ Misses       1550     1518      -32     
+ Partials      426      421       -5     

Flag	Coverage Δ
unittests	73.56% <ø> (+0.43%)	⬆️

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

Impacted Files	Coverage Δ
mmtrack/datasets/pipelines/transforms.py	89.22% <0.00%> (+1.14%)	⬆️
mmtrack/models/trackers/base_tracker.py	86.25% <0.00%> (+1.52%)	⬆️
mmtrack/models/sot/siamrpn.py	81.56% <0.00%> (+1.67%)	⬆️
mmtrack/datasets/pipelines/processing.py	78.44% <0.00%> (+3.44%)	⬆️
mmtrack/datasets/sot_train_dataset.py	85.85% <0.00%> (+8.08%)	⬆️
mmtrack/models/reid/base_reid.py	86.95% <0.00%> (+8.69%)	⬆️
mmtrack/models/trackers/tracktor_tracker.py	79.06% <0.00%> (+15.11%)	⬆️

---

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 503756c...79c66bc. Read the comment docs.