Skip to main content
igodfried
Projects
github_aistream-peelout_flow-forecast
Reports
Effect of long-term dependencies

Effect of long-term dependencies

Isaac
Created on February 7|Last edited on February 11

Comparing training on different sequence lengths




02468epoch0.10.20.30.40.5
02468epoch0.40.60.81
02468epoch200000040000006000000800000010000000
Run set
3
Name
3 visualized
State
Notes
User
Tags
Created
Runtime
GCS
dataset_params.batch_size
dataset_params.class
dataset_params.forecast_history
dataset_params.forecast_length
dataset_params.relevant_cols
dataset_params.scaler
dataset_params.target_col
dataset_params.test_path
dataset_params.train_end
dataset_params.training_path
dataset_params.valid_end
dataset_params.valid_start
dataset_params.validation_path
gcs_m_path_10_model
gcs_m_path_10_params
gcs_m_path_15_model
gcs_m_path_15_params
gcs_m_path_1_model
gcs_m_path_1_params
gcs_path
model_name
model_params.number_time_series
model_params.seq_len
model_type
save_path
training_params.batch_size
training_params.criterion
training_params.epochs
training_params.optim_params.lr
training_params.optimizer
wandb.name
wandb.tags
weight_path
batch_size
dataset_params.feature_param.dataset_params.hour
dataset_params.feature_param.dataset_params.month
dataset_params.feature_param.datetime_param.hour
dataset_params.feature_param.datetime_param.month
dataset_params.feature_param.datetime_params.day
dataset_params.feature_param.datetime_params.day_of_week
dataset_params.feature_param.datetime_params.hour
dataset_params.feature_param.datetime_params.month
dataset_params.feature_params.datetime_param.hour
dataset_params.feature_params.datetime_param.month
Crashed
-
igodfried
01047150
72
SimpleLinear
WVL
corrected
9m 4s
true
200
default
7
1
["cfs1","precip","temp","month"]
StandardScaler
["cfs1"]
day_addition/01047150WVL_flow.csv
35000
day_addition/01047150WVL_flow.csv
40000
35001
day_addition/01047150WVL_flow.csv
gs://predict_cfs/experime...
gs://predict_cfs/experime...
-
-
-
-
-
DARNN
-
-
PyTorch
model_save/30_September_202006_34AM_model.pth
200
MSE
10
0.00001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleLinear","72","corrected"]
-
200
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
72
SimpleLinear
WVL
corrected
1h 26m 31s
true
12
default
48
1
["cfs1","precip","temp","month"]
RobustScaler
["cfs1"]
add_thing2/01047150WVL_flow.csv
35000
add_thing2/01047150WVL_flow.csv
40000
35001
add_thing2/01047150WVL_flow.csv
-
-
-
-
-
-
-
CustomTransformerDecoder
-
-
PyTorch
model_save/20_May_202005_11PM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleLinear","72","corrected"]
model_save/20_May_202005_09PM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
72
SimpleLinear
WVL
corrected
21m 54s
true
12
default
48
1
["cfs1","precip","temp"]
RobustScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
LSTM
-
-
PyTorch
model_save/30_April_202004_49AM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleLinear","72","corrected"]
model_save/30_April_202004_47AM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
72
SimpleLinear
WVL
corrected
1h 32m 13s
true
12
default
48
1
["cfs1","precip","temp"]
RobustScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
CustomTransformerDecoder
-
-
PyTorch
model_save/28_April_202010_41PM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleLinear","72","corrected"]
model_save/28_April_202010_39PM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
72
SimpleLinear
WVL
corrected
1h 24m 7s
true
12
default
48
1
["cfs1","precip","temp"]
RobustScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
CustomTransformerDecoder
-
-
PyTorch
model_save/27_April_202004_13AM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleLinear","72","corrected"]
model_save/27_April_202004_11AM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
51m 36s
true
100
default
50
1
["cfs1","precip","temp"]
RobustScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
model_save/24_April_202005_32AM_model.pth
100
MSE
7
0.0002
SGD
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
model_save/24_April_202005_27AM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
72
SimpleLinear
WVL
corrected
12m 5s
true
100
default
72
1
["cfs1","precip","temp"]
RobustScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleLinearModel
-
-
PyTorch
model_save/22_April_202005_06PM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleLinear","72","corrected"]
model_save/22_April_202005_04PM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
72
SimpleLinear
WVL
corrected
38m 6s
true
100
default
72
1
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleLinearModel
-
-
PyTorch
model_save/21_April_202008_01PM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleLinear","72","corrected"]
model_save/21_April_202008_00PM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
1h 43m 34s
true
100
default
50
1
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
model_save/21_April_202004_37AM_model.pth
100
MSE
7
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
model_save/21_April_202004_32AM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
24
MultiAttnHeadSimple
WVL
corrected
48s
true
100
default
72
2
["cfs1","precip","temp"]
MinMaxScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
MultiAttnHeadSimple
3
72
PyTorch
-
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","MultiAttnHeadSimple","24","corrected"]
-
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
41m 38s
true
100
default
50
1
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
model_save/20_April_202008_59PM_model.pth
100
MSE
7
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
model_save/20_April_202008_54PM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
1h 15m 51s
true
100
default
50
1
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
model_save/20_April_202003_24AM_model.pth
100
MSE
7
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
-
-
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
2m 1s
true
100
default
50
10
["cfs1","precip","temp"]
RobustScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
-
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
-
-
-
-
-
-
-
-
-
-
-
-
Crashed
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
17m 15s
true
100
default
50
10
["cfs1","precip","temp"]
MinMaxScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
-
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
-
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
13m 20s
true
100
default
50
10
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
model_save/17_April_202008_05PM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
-
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
24
MultiAttnHeadSimple
WVL
corrected
34s
true
100
default
48
2
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
MultiAttnHeadSimple
3
48
PyTorch
-
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","MultiAttnHeadSimple","24","corrected"]
-
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
24
MultiAttnHeadSimple
WVL
corrected
16m 15s
true
100
default
24
24
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
MultiAttnHeadSimple
3
24
PyTorch
model_save/08_April_202005_24AM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","MultiAttnHeadSimple","24","corrected"]
model_save/08_April_202005_22AM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
24
MultiAttnHeadSimple
WVL
corrected
16m 52s
true
100
default
24
24
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
MultiAttnHeadSimple
3
24
PyTorch
model_save/06_April_202008_05PM_model.pth
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","MultiAttnHeadSimple","24","corrected"]
model_save/06_April_202008_04PM_model.pth
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
24
MultiAttnHeadSimple
WVL
corrected
20s
true
100
default
24
24
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
MultiAttnHeadSimple
3
24
PyTorch
-
100
MSE
6
0.0001
SGD
flood_forecast_01047150
["01047150","WVL","MultiAttnHeadSimple","24","corrected"]
-
-
-
-
-
-
-
-
-
-
-
-
Finished
-
igodfried
01047150
36
SimpleTransformer
WVL
corrected
43m 26s
true
100
default
100
10
["cfs1","precip","temp"]
StandardScaler
["cfs1"]
joined/01047150WVL_flow.csv
35000
joined/01047150WVL_flow.csv
40000
35001
joined/01047150WVL_flow.csv
-
-
-
-
-
-
-
SimpleTransformer
3
-
PyTorch
model_save/05_April_202009_29PM_model.pth
100
MSE
4
0.0001
Adam
flood_forecast_01047150
["01047150","WVL","SimpleTransformer","36","corrected"]
model_save/05_April_202009_25PM_model.pth
-
-
-
-
-
-
-
-
-
-
-
1-20
of 40


Comparing seq lengths




Run set
19837


Created with ❤️ on Weights & Biases.

https://wandb.ai/igodfried/github_aistream-peelout_flow-forecast/reports/Effect-of-long-term-dependencies--Vmlldzo1MjY4NA
Made with Weights & Biases. Sign up or log in to create reports like this one.