File experiment_meow_main.py changed (mode: 100644) (index 4441b69..0bad71e) |
... |
... |
import os |
17 |
17 |
from model_util import get_lr |
from model_util import get_lr |
18 |
18 |
|
|
19 |
19 |
COMET_ML_API = "S3mM1eMq6NumMxk2QJAXASkUM" |
COMET_ML_API = "S3mM1eMq6NumMxk2QJAXASkUM" |
20 |
|
PROJECT_NAME = "meow-one-experiment-insita" |
|
|
20 |
|
# PROJECT_NAME = "meow-one-experiment-insita" |
|
21 |
|
PROJECT_NAME = "crowd-counting-debug" |
21 |
22 |
|
|
22 |
23 |
|
|
23 |
24 |
def very_simple_param_count(model): |
def very_simple_param_count(model): |
|
... |
... |
if __name__ == "__main__": |
156 |
157 |
experiment.log_metric("batch_timer", batch_timer.value()) |
experiment.log_metric("batch_timer", batch_timer.value()) |
157 |
158 |
experiment.log_metric("train_timer", train_timer.value()) |
experiment.log_metric("train_timer", train_timer.value()) |
158 |
159 |
|
|
|
160 |
|
print("batch_timer ", batch_timer.value()) |
|
161 |
|
print("train_timer ", train_timer.value()) |
|
162 |
|
|
159 |
163 |
@trainer.on(Events.EPOCH_COMPLETED) |
@trainer.on(Events.EPOCH_COMPLETED) |
160 |
164 |
def log_validation_results(trainer): |
def log_validation_results(trainer): |
161 |
165 |
evaluate_timer.resume() |
evaluate_timer.resume() |
|
... |
... |
if __name__ == "__main__": |
173 |
177 |
|
|
174 |
178 |
# timer |
# timer |
175 |
179 |
experiment.log_metric("evaluate_timer", evaluate_timer.value()) |
experiment.log_metric("evaluate_timer", evaluate_timer.value()) |
|
180 |
|
print("evaluate_timer ", evaluate_timer.value()) |
176 |
181 |
|
|
177 |
182 |
def checkpoint_valid_mae_score_function(engine): |
def checkpoint_valid_mae_score_function(engine): |
178 |
183 |
score = engine.state.metrics['valid_mae'] |
score = engine.state.metrics['valid_mae'] |
File local_train_script/M4_t2_shb.sh copied from file train_script/meow_one/M4_t2_shb.sh (similarity 50%) (mode: 100644) (index 6f9663d..d2853c8) |
1 |
|
task="M4_t2_shb" |
|
|
1 |
|
task="local_M4_t2_shb" |
2 |
2 |
|
|
3 |
|
CUDA_VISIBLE_DEVICES=4 HTTPS_PROXY="http://10.60.28.99:86" nohup python experiment_meow_main.py \ |
|
|
3 |
|
nohup python experiment_meow_main.py \ |
4 |
4 |
--task_id $task \ |
--task_id $task \ |
5 |
5 |
--note "M4 shanghaitech_rnd" \ |
--note "M4 shanghaitech_rnd" \ |
6 |
6 |
--model "M4" \ |
--model "M4" \ |
7 |
|
--input /data/rnd/thient/thient_data/ShanghaiTech/part_B \ |
|
|
7 |
|
--input /data/ShanghaiTech/part_B \ |
8 |
8 |
--lr 1e-4 \ |
--lr 1e-4 \ |
9 |
9 |
--decay 1e-4 \ |
--decay 1e-4 \ |
10 |
|
--batch_size 8 \ |
|
|
10 |
|
--batch_size 5 \ |
11 |
11 |
--datasetname shanghaitech_rnd \ |
--datasetname shanghaitech_rnd \ |
12 |
12 |
--epochs 301 > logs/$task.log & |
--epochs 301 > logs/$task.log & |
13 |
13 |
|
|