-
Notifications
You must be signed in to change notification settings - Fork 0
/
ReadMe
executable file
·44 lines (35 loc) · 2.17 KB
/
ReadMe
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
1. PreProcessing:
1.a (vir) python process_main.py [INPUT File in Data Folder] [SPECIFY OUTPUT FOLDER (autocreate in Results/)] [PARAMETERS]
- e.g. python process_main.py reviews_small.txt reviews_small_output_test PDTOWC
2. Create Datasets: (openke)
2.a python make_dataset.py [ INPUT FOLDER PATH ] [INPUT FOLDER]
- Output Folder will be created in OpenKE/Dataset depending on input Folder
- Two Folder will be create: With and Without Co-occurence
- e.g. python make_dataset.py ~/code/Final\ Code/Result Reviews_NEW_OUTPUT_11
- e.g. python make_dataset.py ~/code/Final\ Code/Result Reviews_V10K
2.b python analysis.py Dataset/[FOLDER]
- Count Relation-wise, Head-wise and Tail-wise.
- create file analysis.txt for Results
- e.g. python analysis.py Dataset/Reviews_V10K_WN_DP > Dataset/Reviews_V10K_WN_DP/analysis.txt
2.c python graph_analysis.py Dataset/[FOLDER]
- Relation and Entity Sorted by Vocab, Count and ID and Find unique count.
- entity/relation/train_stats.txt will be created
- e.g.python graph_analysis.py Dataset/Reviews_V10K_WN_DP
2.d python remove_duplicate.py Dataset/[FOLDER]
- Create Duplicated and Unique limited relation.
- Created folder Dataset/[FOLDER]_Unique
- After this you can run analysis.py and graph_analysis.py for more info
- e.g. python remove_duplicate.py Dataset/Reviews_V10K_WN_DP
2.e Useful DP
- python remove_dp_relations.py [Folder]
- remove_dp_relations.py ===>>> useful relation.
3. Training: (openke)
3.a python train.py python train.py [INPUT FOLDER in Dataset] [OUTPUT FOLDER in res] [loss_file_modifier]
- python train.py Reviews_NEW_OUTPUT_11_WN_DP Reviews_OUT_Adadelta_E1000_L1_D150_B32 0
4. Evaluation:
4.c (openke) python convert_vector.py [Dataset] [FOLDER]
- python convert_vector.py Reviews_All_Rels Reviews_Adadelta_E5000_B32_D150_L1
- create embedding vector for processing
4.b (vir) python evaluate.py [path to embedding vector] [output_modifier]
- Evaluate learnt embedding based on various Evaluation.
- create file OUT_out_modifier