Informer模型实战案例(代码+数据集+参数讲解)

资源文件列表(大概)

文件名

大小

.idea/

.idea/.gitignore

50B

.idea/aws.xml

304B

.idea/Informer2020-main.iml

482B

.idea/inspectionProfiles/

.idea/inspectionProfiles/profiles_settings.xml

174B

.idea/inspectionProfiles/Project_Default.xml

1.27KB

.idea/misc.xml

185B

.idea/modules.xml

293B

.idea/workspace.xml

6.02KB

checkpoints/

checkpoints/informer_custom_ftMS_sl126_ll64_pl24_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/

checkpoints/informer_custom_ftMS_sl126_ll64_pl24_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/checkpoint.pth

62.77MB

checkpoints/informer_custom_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/

checkpoints/informer_custom_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/checkpoint.pth

62.77MB

data/

data/__init__.py

data/__pycache__/

data/__pycache__/__init__.cpython-39.pyc

153B

data/__pycache__/data_loader.cpython-39.pyc

8.79KB

data/data_loader.py

13.42KB

environment.yml

198B

ETTh1-Test.csv

38.37KB

ETTh1.csv

2.47MB

exp/

exp/__init__.py

exp/__pycache__/

exp/__pycache__/__init__.cpython-39.pyc

152B

exp/__pycache__/exp_basic.cpython-39.pyc

1.54KB

exp/__pycache__/exp_informer.cpython-39.pyc

7.75KB

exp/exp_basic.py

875B

exp/exp_informer.py

10.83KB

forecsat.csv

265B

main_informer.py

7.26KB

models/

models/__init__.py

models/__pycache__/

models/__pycache__/__init__.cpython-39.pyc

155B

models/__pycache__/attn.cpython-39.pyc

5.02KB

models/__pycache__/decoder.cpython-39.pyc

1.95KB

models/__pycache__/embed.cpython-39.pyc

5.04KB

models/__pycache__/encoder.cpython-39.pyc

3.49KB

models/__pycache__/model.cpython-39.pyc

4.78KB

models/attn.py

6.03KB

models/decoder.py

1.73KB

models/embed.py

4.04KB

models/encoder.py

3.47KB

models/model.py

7.05KB

results/

results/informer_custom_ftMS_sl126_ll64_pl24_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/

results/informer_custom_ftMS_sl126_ll64_pl24_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/metrics.npy

148B

results/informer_custom_ftMS_sl126_ll64_pl24_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/pred.npy

12.13KB

results/informer_custom_ftMS_sl126_ll64_pl24_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/true.npy

12.13KB

results/informer_custom_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/

results/informer_custom_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/metrics.npy

148B

results/informer_custom_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/pred.npy

2.13KB

results/informer_custom_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/true.npy

2.13KB

results/informer_Sum_ftM_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/

results/informer_Sum_ftM_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/metrics.npy

148B

results/informer_Sum_ftM_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/pred.npy

27.13KB

results/informer_Sum_ftM_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/real_prediction.npy

176B

results/informer_Sum_ftM_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/true.npy

27.13KB

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/metrics.npy

148B

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/pred.npy

9.13KB

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/real_prediction.npy

144B

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_0/true.npy

9.13KB

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_1/

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_1/metrics.npy

148B

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_1/pred.npy

2.13KB

results/informer_Sum_ftMS_sl126_ll64_pl4_dm512_nh8_el2_dl1_df2048_atprob_fc5_ebtimeF_dtTrue_mxTrue_test_1/true.npy

2.13KB

utils/

utils/__init__.py

utils/__pycache__/

utils/__pycache__/__init__.cpython-39.pyc

154B

utils/__pycache__/masking.cpython-39.pyc

1.43KB

utils/__pycache__/metrics.cpython-39.pyc

1.42KB

utils/__pycache__/timefeatures.cpython-39.pyc

7.16KB

utils/__pycache__/tools.cpython-39.pyc

3.21KB

utils/masking.py

851B

utils/metrics.py

826B

utils/timefeatures.py

5.43KB

utils/tools.py

2.76KB

资源内容介绍

本篇博客带大家看的是Informer模型进行时间序列预测的实战案例，它是在2019年被提出并在ICLR 2020上被评为Best Paper，可以说Informer模型在当今的时间序列预测方面还是十分可靠的，Informer模型的实质是注意力机制+Transformer模型，Informer模型的核心思想是将输入序列进行自注意力机制的处理，以捕捉序列中的长期依赖关系，并利用Transformer的编码器-解码器结构进行预测，通过阅读本文你可以学会利用个人数据集训练模型。Informer是一种用于长序列时间序列预测的Transformer模型,但是它与传统的Transformer模型又有些不同点，与传统的Transformer模型相比，Informer具有以下几个独特的特点：1. ProbSparse自注意力机制：Informer引入了ProbSparse自注意力机制，该机制在时间复杂度和内存使用方面达到了O(Llog L)的水平，能够有效地捕捉序列之间的长期依赖关系。2. 自注意力蒸馏：通过减少级联层的输入，自注意力蒸馏技术可以有效处理极长的输入序列，提高了模型处理长序列的能力

用户评论 (0)

发表评论

您的昵称:

评分

评论内容:

验证码:

Informer模型实战案例(代码+数据集+参数讲解)

资源文件列表(大概)

资源内容介绍

用户评论 (0)

发表评论

相关资源