try sep vit

data/icews14_both/about.txt

@@ -0,0 +1,15 @@
+# triples: 86517 
+# entities: 7128 
+# relations: 12409 
+# timesteps: 208 
+# test triples: 8218 
+# valid triples: 8193 
+# train triples: 70106 
+Measure method:  N/A  
+Target Size :  0  
+Grow Factor:  0  
+Shrink Factor:  0  
+Epsilon Factor: 0  
+Search method: N/A  
+filter_dupes: both
+nonames: False

data/icews14_both/time_map.dict

@@ -0,0 +1,209 @@
+0	0	2
+1	3	5
+2	6	7
+3	8	9
+4	10	12
+5	13	14
+6	15	16
+7	17	19
+8	20	21
+9	22	23
+10	24	26
+11	27	28
+12	29	30
+13	31	33
+14	34	35
+15	36	37
+16	38	40
+17	41	42
+18	43	44
+19	45	46
+20	47	48
+21	49	49
+22	50	50
+23	51	51
+24	52	53
+25	54	54
+26	55	55
+27	56	57
+28	58	59
+29	60	61
+30	62	62
+31	63	63
+32	64	65
+33	66	68
+34	69	70
+35	71	71
+36	72	72
+37	73	74
+38	75	76
+39	77	78
+40	79	80
+41	81	82
+42	83	84
+43	85	85
+44	86	87
+45	88	89
+46	90	91
+47	92	93
+48	94	96
+49	97	97
+50	98	99
+51	100	101
+52	102	103
+53	104	105
+54	106	107
+55	108	110
+56	111	112
+57	113	114
+58	115	116
+59	117	118
+60	119	119
+61	120	121
+62	122	124
+63	125	125
+64	126	127
+65	128	129
+66	130	131
+67	132	133
+68	134	135
+69	136	138
+70	139	139
+71	140	140
+72	141	141
+73	142	143
+74	144	145
+75	146	147
+76	148	148
+77	149	150
+78	151	152
+79	153	154
+80	155	155
+81	156	157
+82	158	159
+83	160	161
+84	162	163
+85	164	166
+86	167	167
+87	168	168
+88	169	169
+89	170	170
+90	171	173
+91	174	175
+92	176	177
+93	178	180
+94	181	182
+95	183	183
+96	184	185
+97	186	187
+98	188	188
+99	189	190
+100	191	192
+101	193	194
+102	195	195
+103	196	197
+104	198	199
+105	200	201
+106	202	203
+107	204	205
+108	206	208
+109	209	210
+110	211	212
+111	213	215
+112	216	217
+113	218	219
+114	220	221
+115	222	222
+116	223	224
+117	225	226
+118	227	229
+119	230	231
+120	232	233
+121	234	236
+122	237	238
+123	239	239
+124	240	241
+125	242	243
+126	244	245
+127	246	246
+128	247	248
+129	249	250
+130	251	251
+131	252	252
+132	253	253
+133	254	254
+134	255	256
+135	257	257
+136	258	259
+137	260	261
+138	262	263
+139	264	264
+140	265	265
+141	266	266
+142	267	267
+143	268	269
+144	270	271
+145	272	272
+146	273	273
+147	274	274
+148	275	276
+149	277	278
+150	279	279
+151	280	281
+152	282	283
+153	284	285
+154	286	286
+155	287	287
+156	288	288
+157	289	289
+158	290	291
+159	292	292
+160	293	293
+161	294	294
+162	295	295
+163	296	297
+164	298	299
+165	300	300
+166	301	301
+167	302	303
+168	304	305
+169	306	307
+170	308	309
+171	310	310
+172	311	312
+173	313	313
+174	314	314
+175	315	315
+176	316	316
+177	317	317
+178	318	319
+179	320	320
+180	321	321
+181	322	322
+182	323	323
+183	324	324
+184	325	326
+185	327	327
+186	328	328
+187	329	329
+188	330	330
+189	331	332
+190	333	334
+191	335	335
+192	336	336
+193	337	338
+194	339	340
+195	341	342
+196	343	343
+197	344	344
+198	345	346
+199	347	348
+200	349	349
+201	350	350
+202	351	352
+203	353	355
+204	356	357
+205	358	359
+206	360	362
+207	363	365
+208	366	366

data/wikidata12k_both/about.txt

@@ -0,0 +1,15 @@
+# triples: 231529 
+# entities: 12554 
+# relations: 423 
+# timesteps: 70 
+# test triples: 16195 
+# valid triples: 16707 
+# train triples: 198627 
+Measure method:  N/A  
+Target Size :  423  
+Grow Factor:  0  
+Shrink Factor:  4.0  
+Epsilon Factor: 0  
+Search method: N/A  
+filter_dupes: both
+nonames: False

data/wikidata12k_both/relations.dict

@@ -0,0 +1,423 @@
+0	P131[0-0]
+1	P131[1-1]
+2	P131[2-2]
+3	P131[3-3]
+4	P131[4-4]
+5	P131[5-5]
+6	P131[6-6]
+7	P131[7-7]
+8	P131[8-8]
+9	P131[9-9]
+10	P131[10-10]
+11	P131[11-11]
+12	P131[12-12]
+13	P131[13-13]
+14	P131[14-14]
+15	P131[15-15]
+16	P131[16-16]
+17	P131[17-17]
+18	P131[18-18]
+19	P131[19-19]
+20	P131[20-20]
+21	P131[21-21]
+22	P131[22-22]
+23	P131[23-23]
+24	P131[24-24]
+25	P131[25-25]
+26	P131[26-26]
+27	P131[27-27]
+28	P131[28-28]
+29	P131[29-29]
+30	P131[30-30]
+31	P131[31-31]
+32	P131[32-32]
+33	P131[33-33]
+34	P131[34-34]
+35	P131[35-35]
+36	P131[36-36]
+37	P131[37-37]
+38	P131[38-38]
+39	P131[39-39]
+40	P131[40-40]
+41	P131[41-41]
+42	P131[42-42]
+43	P131[43-43]
+44	P131[44-44]
+45	P131[45-45]
+46	P131[46-46]
+47	P131[47-47]
+48	P131[48-48]
+49	P131[49-49]
+50	P131[50-50]
+51	P131[51-51]
+52	P131[52-52]
+53	P131[53-53]
+54	P131[54-54]
+55	P131[55-55]
+56	P131[56-56]
+57	P131[57-57]
+58	P131[58-58]
+59	P131[59-59]
+60	P131[60-60]
+61	P131[61-61]
+62	P131[62-62]
+63	P131[63-63]
+64	P131[64-64]
+65	P131[65-65]
+66	P131[66-66]
+67	P131[67-67]
+68	P131[68-68]
+69	P131[69-69]
+70	P1435[65-65]
+71	P39[49-49]
+72	P39[50-50]
+73	P39[51-51]
+74	P39[52-52]
+75	P39[53-53]
+76	P39[54-54]
+77	P39[55-55]
+78	P39[56-56]
+79	P39[57-57]
+80	P39[58-58]
+81	P39[59-59]
+82	P39[60-60]
+83	P39[61-61]
+84	P39[62-62]
+85	P39[63-63]
+86	P39[64-64]
+87	P39[65-65]
+88	P39[66-66]
+89	P39[67-67]
+90	P39[68-68]
+91	P39[69-69]
+92	P54[40-40]
+93	P54[41-41]
+94	P54[42-42]
+95	P54[43-43]
+96	P54[44-44]
+97	P54[45-45]
+98	P54[46-46]
+99	P54[47-47]
+100	P54[48-48]
+101	P54[49-49]
+102	P54[50-50]
+103	P54[51-51]
+104	P54[52-52]
+105	P54[53-53]
+106	P54[54-54]
+107	P54[55-55]
+108	P54[56-56]
+109	P54[57-57]
+110	P54[58-58]
+111	P54[59-59]
+112	P54[60-60]
+113	P54[61-61]
+114	P54[62-62]
+115	P54[63-63]
+116	P54[64-64]
+117	P54[65-65]
+118	P54[66-66]
+119	P54[67-67]
+120	P54[68-68]
+121	P54[69-69]
+122	P31[0-0]
+123	P31[1-1]
+124	P31[2-2]
+125	P31[3-3]
+126	P31[4-4]
+127	P31[5-5]
+128	P31[6-6]
+129	P31[7-7]
+130	P31[8-8]
+131	P31[9-9]
+132	P31[10-10]
+133	P31[11-11]
+134	P31[12-12]
+135	P31[13-13]
+136	P31[14-14]
+137	P31[15-15]
+138	P31[16-16]
+139	P31[17-17]
+140	P31[18-18]
+141	P31[19-19]
+142	P31[20-20]
+143	P31[21-21]
+144	P31[22-22]
+145	P31[23-23]
+146	P31[24-24]
+147	P31[25-25]
+148	P31[26-26]
+149	P31[27-27]
+150	P31[28-28]
+151	P31[29-29]
+152	P31[30-30]
+153	P31[31-31]
+154	P31[32-32]
+155	P31[33-33]
+156	P31[34-34]
+157	P31[35-35]
+158	P31[36-36]
+159	P31[37-37]
+160	P31[38-38]
+161	P31[39-39]
+162	P31[40-40]
+163	P31[41-41]
+164	P31[42-42]
+165	P31[43-43]
+166	P31[44-44]
+167	P31[45-45]
+168	P31[46-46]
+169	P31[47-47]
+170	P31[48-48]
+171	P31[49-49]
+172	P31[50-50]
+173	P31[51-51]
+174	P31[52-52]
+175	P31[53-53]
+176	P31[54-54]
+177	P31[55-55]
+178	P31[56-56]
+179	P31[57-57]
+180	P31[58-58]
+181	P31[59-59]
+182	P31[60-60]
+183	P31[61-61]
+184	P31[62-62]
+185	P31[63-63]
+186	P31[64-64]
+187	P31[65-65]
+188	P31[66-66]
+189	P31[67-67]
+190	P31[68-68]
+191	P31[69-69]
+192	P463[26-26]
+193	P463[27-27]
+194	P463[28-28]
+195	P463[29-29]
+196	P463[30-30]
+197	P463[31-31]
+198	P463[32-32]
+199	P463[33-33]
+200	P463[34-34]
+201	P463[35-35]
+202	P463[36-36]
+203	P463[37-37]
+204	P463[38-38]
+205	P463[39-39]
+206	P463[40-40]
+207	P463[41-41]
+208	P463[42-42]
+209	P463[43-43]
+210	P463[44-44]
+211	P463[45-45]
+212	P463[46-46]
+213	P463[47-47]
+214	P463[48-48]
+215	P463[49-49]
+216	P463[50-50]
+217	P463[51-51]
+218	P463[52-52]
+219	P463[53-53]
+220	P463[54-54]
+221	P463[55-55]
+222	P463[56-56]
+223	P463[57-57]
+224	P463[58-58]
+225	P463[59-59]
+226	P463[60-60]
+227	P463[61-61]
+228	P463[62-62]
+229	P463[63-63]
+230	P463[64-64]
+231	P463[65-65]
+232	P463[66-66]
+233	P463[67-67]
+234	P463[68-68]
+235	P463[69-69]
+236	P512[4-69]
+237	P190[0-29]
+238	P150[0-3]
+239	P1376[39-47]
+240	P463[0-7]
+241	P166[0-7]
+242	P2962[18-30]
+243	P108[29-36]
+244	P39[0-3]
+245	P17[47-48]
+246	P166[21-23]
+247	P793[46-69]
+248	P69[32-41]
+249	P17[57-58]
+250	P190[42-45]
+251	P2962[39-42]
+252	P54[0-18]
+253	P26[56-61]
+254	P150[14-17]
+255	P463[16-17]
+256	P26[39-46]
+257	P579[36-43]
+258	P579[16-23]
+259	P2962[59-60]
+260	P1411[59-61]
+261	P26[20-27]
+262	P6[4-69]
+263	P1435[33-34]
+264	P166[52-53]
+265	P108[49-57]
+266	P150[10-13]
+267	P1346[47-68]
+268	P150[18-21]
+269	P1346[13-46]
+270	P69[20-23]
+271	P39[31-32]
+272	P1411[32-37]
+273	P166[62-63]
+274	P150[44-47]
+275	P2962[61-62]
+276	P150[48-51]
+277	P150[52-55]
+278	P1411[62-67]
+279	P1435[35-36]
+280	P1411[48-51]
+281	P150[22-25]
+282	P2962[63-64]
+283	P2962[65-66]
+284	P166[58-59]
+285	P190[46-49]
+286	P54[34-35]
+287	P1435[4-16]
+288	P463[18-19]
+289	P150[31-34]
+290	P150[35-38]
+291	P39[35-36]
+292	P26[62-69]
+293	P1411[56-58]
+294	P1435[37-38]
+295	P166[60-61]
+296	P39[33-34]
+297	P102[24-31]
+298	P2962[43-46]
+299	P108[37-48]
+300	P190[50-53]
+301	P39[4-6]
+302	P1435[39-40]
+303	P793[0-45]
+304	P150[64-69]
+305	P39[19-22]
+306	P27[30-38]
+307	P2962[31-38]
+308	P1411[24-31]
+309	P102[40-45]
+310	P39[37-38]
+311	P463[8-11]
+312	P1435[41-42]
+313	P27[52-59]
+314	P69[16-19]
+315	P17[16-18]
+316	P190[54-57]
+317	P1435[43-44]
+318	P166[8-15]
+319	P166[45-47]
+320	P2962[47-50]
+321	P39[39-40]
+322	P1411[52-55]
+323	P108[58-69]
+324	P463[20-21]
+325	P39[41-42]
+326	P150[26-30]
+327	P150[39-43]
+328	P1435[45-46]
+329	P26[28-38]
+330	P54[27-30]
+331	P190[58-61]
+332	P17[59-61]
+333	P54[36-37]
+334	P166[16-20]
+335	P166[37-40]
+336	P1435[47-48]
+337	P17[0-3]
+338	P26[47-55]
+339	P1435[49-50]
+340	P1435[25-28]
+341	P150[4-9]
+342	P102[63-69]
+343	P26[0-19]
+344	P1435[17-24]
+345	P39[23-26]
+346	P1435[51-52]
+347	P39[7-11]
+348	P69[12-15]
+349	P69[24-31]
+350	P102[0-23]
+351	P39[43-44]
+352	P579[24-35]
+353	P190[62-65]
+354	P1435[53-54]
+355	P1376[0-18]
+356	P27[0-14]
+357	P463[12-15]
+358	P166[33-36]
+359	P102[32-39]
+360	P17[4-7]
+361	P190[30-41]
+362	P166[24-28]
+363	P190[66-69]
+364	P69[42-69]
+365	P1435[55-56]
+366	P54[31-33]
+367	P39[45-46]
+368	P17[12-15]
+369	P1435[57-58]
+370	P54[19-26]
+371	P2962[51-54]
+372	P2962[67-69]
+373	P1435[59-60]
+374	P579[44-56]
+375	P1435[61-62]
+376	P166[41-44]
+377	P17[19-22]
+378	P1376[19-38]
+379	P17[23-26]
+380	P1376[48-69]
+381	P463[22-23]
+382	P17[27-30]
+383	P1435[63-64]
+384	P69[0-3]
+385	P1435[66-67]
+386	P17[35-38]
+387	P69[8-11]
+388	P1435[68-69]
+389	P17[31-34]
+390	P102[46-53]
+391	P27[60-69]
+392	P579[57-69]
+393	P69[4-7]
+394	P1411[7-14]
+395	P551[0-35]
+396	P108[0-28]
+397	P17[8-11]
+398	P1411[38-47]
+399	P17[43-46]
+400	P17[49-52]
+401	P166[64-69]
+402	P1435[29-32]
+403	P54[38-39]
+404	P39[27-30]
+405	P2962[55-58]
+406	P463[24-25]
+407	P17[39-42]
+408	P17[53-56]
+409	P17[66-69]
+410	P17[62-65]
+411	P1411[15-23]
+412	P166[48-51]
+413	P27[15-29]
+414	P150[56-63]
+415	P27[39-51]
+416	P39[47-48]
+417	P166[29-32]
+418	P39[12-18]
+419	P166[54-57]
+420	P551[36-69]
+421	P579[0-15]
+422	P102[54-62]

data/wikidata12k_both/time_map.dict

@@ -0,0 +1,71 @@
+0	19	19
+1	20	1643
+2	1644	1790
+3	1791	1816
+4	1817	1855
+5	1856	1871
+6	1872	1893
+7	1894	1905
+8	1906	1913
+9	1914	1918
+10	1919	1920
+11	1921	1924
+12	1925	1929
+13	1930	1933
+14	1934	1937
+15	1938	1941
+16	1942	1945
+17	1946	1948
+18	1949	1950
+19	1951	1953
+20	1954	1956
+21	1957	1959
+22	1960	1961
+23	1962	1963
+24	1964	1965
+25	1966	1967
+26	1968	1968
+27	1969	1970
+28	1971	1972
+29	1973	1974
+30	1975	1976
+31	1977	1978
+32	1979	1980
+33	1981	1982
+34	1983	1983
+35	1984	1984
+36	1985	1985
+37	1986	1986
+38	1987	1987
+39	1988	1988
+40	1989	1989
+41	1990	1990
+42	1991	1991
+43	1992	1992
+44	1993	1993
+45	1994	1994
+46	1995	1995
+47	1996	1996
+48	1997	1997
+49	1998	1998
+50	1999	1999
+51	2000	2000
+52	2001	2001
+53	2002	2002
+54	2003	2003
+55	2004	2004
+56	2005	2005
+57	2006	2006
+58	2007	2007
+59	2008	2008
+60	2009	2009
+61	2010	2010
+62	2011	2011
+63	2012	2012
+64	2013	2013
+65	2014	2014
+66	2015	2015
+67	2016	2016
+68	2017	2017
+69	2018	2020
+70	2021	2021

data/yago11k_both/about.txt

@@ -0,0 +1,15 @@
+# triples: 78032 
+# entities: 10526 
+# relations: 177 
+# timesteps: 46 
+# test triples: 6909 
+# valid triples: 7198 
+# train triples: 63925 
+Measure method:  N/A  
+Target Size :  0  
+Grow Factor:  0  
+Shrink Factor:  0  
+Epsilon Factor: 5.0  
+Search method: N/A  
+filter_dupes: both
+nonames: False

data/yago11k_both/relations.dict

@@ -0,0 +1,177 @@
+0	<wasBornIn>[0-2]
+1	<wasBornIn>[2-5]
+2	<wasBornIn>[5-7]
+3	<wasBornIn>[7-10]
+4	<wasBornIn>[10-12]
+5	<wasBornIn>[12-15]
+6	<wasBornIn>[15-17]
+7	<wasBornIn>[17-20]
+8	<wasBornIn>[20-22]
+9	<wasBornIn>[22-25]
+10	<wasBornIn>[25-27]
+11	<wasBornIn>[27-30]
+12	<wasBornIn>[30-32]
+13	<wasBornIn>[32-35]
+14	<wasBornIn>[35-45]
+15	<wasBornIn>[52-52]
+16	<diedIn>[0-3]
+17	<diedIn>[3-5]
+18	<diedIn>[5-7]
+19	<diedIn>[7-10]
+20	<diedIn>[10-12]
+21	<diedIn>[12-14]
+22	<diedIn>[14-17]
+23	<diedIn>[17-19]
+24	<diedIn>[19-21]
+25	<diedIn>[21-23]
+26	<diedIn>[23-25]
+27	<diedIn>[25-27]
+28	<diedIn>[27-29]
+29	<diedIn>[29-32]
+30	<diedIn>[32-34]
+31	<diedIn>[34-36]
+32	<diedIn>[36-38]
+33	<diedIn>[38-40]
+34	<diedIn>[40-42]
+35	<diedIn>[42-44]
+36	<diedIn>[44-47]
+37	<diedIn>[47-49]
+38	<diedIn>[49-51]
+39	<diedIn>[51-53]
+40	<diedIn>[53-55]
+41	<diedIn>[55-57]
+42	<diedIn>[59-59]
+43	<worksAt>[0-3]
+44	<worksAt>[3-5]
+45	<worksAt>[5-7]
+46	<worksAt>[7-10]
+47	<worksAt>[10-12]
+48	<worksAt>[12-14]
+49	<worksAt>[14-17]
+50	<worksAt>[17-19]
+51	<worksAt>[19-21]
+52	<worksAt>[21-23]
+53	<worksAt>[23-25]
+54	<worksAt>[25-27]
+55	<worksAt>[27-29]
+56	<worksAt>[29-32]
+57	<worksAt>[32-34]
+58	<worksAt>[34-36]
+59	<worksAt>[36-40]
+60	<worksAt>[40-42]
+61	<worksAt>[42-47]
+62	<worksAt>[47-53]
+63	<worksAt>[59-59]
+64	<playsFor>[0-3]
+65	<playsFor>[3-5]
+66	<playsFor>[5-23]
+67	<playsFor>[23-25]
+68	<playsFor>[25-27]
+69	<playsFor>[27-29]
+70	<playsFor>[29-32]
+71	<playsFor>[32-34]
+72	<playsFor>[34-36]
+73	<playsFor>[36-38]
+74	<playsFor>[38-40]
+75	<playsFor>[40-42]
+76	<playsFor>[42-44]
+77	<playsFor>[44-47]
+78	<playsFor>[47-51]
+79	<playsFor>[59-59]
+80	<hasWonPrize>[1-4]
+81	<hasWonPrize>[4-6]
+82	<hasWonPrize>[6-8]
+83	<hasWonPrize>[8-11]
+84	<hasWonPrize>[11-15]
+85	<hasWonPrize>[15-18]
+86	<hasWonPrize>[18-22]
+87	<hasWonPrize>[22-26]
+88	<hasWonPrize>[26-30]
+89	<hasWonPrize>[30-33]
+90	<hasWonPrize>[33-37]
+91	<hasWonPrize>[37-47]
+92	<hasWonPrize>[47-53]
+93	<hasWonPrize>[59-59]
+94	<isMarriedTo>[0-3]
+95	<isMarriedTo>[3-5]
+96	<isMarriedTo>[5-7]
+97	<isMarriedTo>[7-10]
+98	<isMarriedTo>[10-12]
+99	<isMarriedTo>[12-14]
+100	<isMarriedTo>[14-17]
+101	<isMarriedTo>[17-19]
+102	<isMarriedTo>[19-21]
+103	<isMarriedTo>[21-23]
+104	<isMarriedTo>[23-25]
+105	<isMarriedTo>[25-27]
+106	<isMarriedTo>[27-29]
+107	<isMarriedTo>[29-32]
+108	<isMarriedTo>[32-34]
+109	<isMarriedTo>[34-38]
+110	<isMarriedTo>[38-42]
+111	<isMarriedTo>[42-47]
+112	<isMarriedTo>[47-51]
+113	<isMarriedTo>[51-55]
+114	<isMarriedTo>[59-59]
+115	<owns>[0-10]
+116	<owns>[10-17]
+117	<owns>[17-19]
+118	<owns>[19-23]
+119	<owns>[23-36]
+120	<owns>[36-38]
+121	<owns>[59-59]
+122	<graduatedFrom>[0-3]
+123	<graduatedFrom>[3-5]
+124	<graduatedFrom>[5-7]
+125	<graduatedFrom>[7-10]
+126	<graduatedFrom>[10-14]
+127	<graduatedFrom>[14-17]
+128	<graduatedFrom>[17-19]
+129	<graduatedFrom>[19-21]
+130	<graduatedFrom>[21-23]
+131	<graduatedFrom>[23-27]
+132	<graduatedFrom>[27-32]
+133	<graduatedFrom>[32-34]
+134	<graduatedFrom>[34-38]
+135	<graduatedFrom>[38-42]
+136	<graduatedFrom>[59-59]
+137	<isAffiliatedTo>[1-4]
+138	<isAffiliatedTo>[4-6]
+139	<isAffiliatedTo>[6-8]
+140	<isAffiliatedTo>[8-11]
+141	<isAffiliatedTo>[11-13]
+142	<isAffiliatedTo>[13-15]
+143	<isAffiliatedTo>[15-18]
+144	<isAffiliatedTo>[18-20]
+145	<isAffiliatedTo>[20-22]
+146	<isAffiliatedTo>[22-24]
+147	<isAffiliatedTo>[24-26]
+148	<isAffiliatedTo>[26-28]
+149	<isAffiliatedTo>[28-30]
+150	<isAffiliatedTo>[30-33]
+151	<isAffiliatedTo>[33-35]
+152	<isAffiliatedTo>[35-37]
+153	<isAffiliatedTo>[37-40]
+154	<isAffiliatedTo>[40-42]
+155	<isAffiliatedTo>[42-44]
+156	<isAffiliatedTo>[44-47]
+157	<isAffiliatedTo>[47-49]
+158	<isAffiliatedTo>[49-51]
+159	<isAffiliatedTo>[51-53]
+160	<isAffiliatedTo>[53-55]
+161	<isAffiliatedTo>[55-57]
+162	<isAffiliatedTo>[59-59]
+163	<created>[0-3]
+164	<created>[3-5]
+165	<created>[5-10]
+166	<created>[10-12]
+167	<created>[12-17]
+168	<created>[17-19]
+169	<created>[19-25]
+170	<created>[25-29]
+171	<created>[29-32]
+172	<created>[32-36]
+173	<created>[36-42]
+174	<created>[42-47]
+175	<created>[47-53]
+176	<created>[59-59]

data/yago11k_both/time_map.dict

@@ -0,0 +1,60 @@
+0	-431	1782
+1	1783	1848
+2	1849	1870
+3	1871	1888
+4	1889	1899
+5	1900	1906
+6	1907	1912
+7	1913	1917
+8	1918	1922
+9	1923	1926
+10	1927	1930
+11	1931	1934
+12	1935	1938
+13	1939	1941
+14	1942	1944
+15	1945	1947
+16	1948	1950
+17	1951	1953
+18	1954	1956
+19	1957	1959
+20	1960	1962
+21	1963	1965
+22	1966	1967
+23	1968	1969
+24	1970	1971
+25	1972	1973
+26	1974	1975
+27	1976	1977
+28	1978	1979
+29	1980	1981
+30	1982	1983
+31	1984	1985
+32	1986	1987
+33	1988	1989
+34	1990	1991
+35	1992	1993
+36	1994	1994
+37	1995	1996
+38	1997	1997
+39	1998	1998
+40	1999	1999
+41	2000	2000
+42	2001	2001
+43	2002	2002
+44	2003	2003
+45	2004	2004
+46	2005	2005
+47	2006	2006
+48	2007	2007
+49	2008	2008
+50	2009	2009
+51	2010	2010
+52	2011	2011
+53	2012	2012
+54	2013	2013
+55	2014	2014
+56	2015	2015
+57	2016	2016
+58	2017	2017
+59	2018	2018

icews14_both.log

@@ -0,0 +1,3 @@
+nohup: ignoring input
+2023-06-20 09:22:51,618 - [INFO] - {'dataset': 'icews14_both', 'name': 'icews14_both', 'gpu': '2', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': False}
+2023-06-20 09:22:57,979 - [INFO] - [E:0| 0]: Train Loss:0.70005,  Val MRR:0.0, 	icews14_both

log/fb_one_to_x

@@ -1964,3 +1964,25 @@
 2023-05-04 08:27:31,384 - fb_one_to_x - [INFO] - [E:34| 1500]: Train Loss:0.0027362,  Val MRR:0.33574, 	fb_one_to_x
 2023-05-04 08:29:20,404 - fb_one_to_x - [INFO] - [E:34| 1600]: Train Loss:0.0027362,  Val MRR:0.33574, 	fb_one_to_x
 2023-05-04 08:31:12,139 - fb_one_to_x - [INFO] - [E:34| 1700]: Train Loss:0.0027362,  Val MRR:0.33574, 	fb_one_to_x
+2023-05-04 08:55:56,065 - fb_one_to_x - [INFO] - {'dataset': 'FB15k-237', 'name': 'fb_one_to_x', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.1, 'drop': 0.2, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': True}
+2023-05-04 08:56:07,953 - fb_one_to_x - [INFO] - [Test, Tail_Batch Step 0]	fb_one_to_x
+2023-05-04 08:56:53,173 - fb_one_to_x - [INFO] - [Test, Tail_Batch Step 100]	fb_one_to_x
+2023-05-04 08:57:20,187 - fb_one_to_x - [INFO] - [Test, Head_Batch Step 0]	fb_one_to_x
+2023-05-04 08:58:08,090 - fb_one_to_x - [INFO] - [Test, Head_Batch Step 100]	fb_one_to_x
+2023-05-04 08:58:36,338 - fb_one_to_x - [INFO] - [Evaluating Epoch 0 test]: 
+	MRR: Tail : 0.43029, Head : 0.23256, Avg : 0.33142
+	MR: Tail : 149.91, Head : 288.48, Avg : 219.2
+	Hit-1: Tail : 0.33563, Head : 0.14883, Avg : 0.24223
+	Hit-3: Tail : 0.47068, Head : 0.25515, Avg : 0.36292
+	Hit-10: Tail : 0.61952, Head : 0.40096, Avg : 0.51024
+2023-05-04 09:03:55,555 - fb_one_to_x - [INFO] - {'dataset': 'FB15k-237', 'name': 'fb_one_to_x', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.1, 'drop': 0.2, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': True}
+2023-05-04 09:04:07,491 - fb_one_to_x - [INFO] - [Test, Tail_Batch Step 0]	fb_one_to_x
+2023-05-04 09:04:52,620 - fb_one_to_x - [INFO] - [Test, Tail_Batch Step 100]	fb_one_to_x
+2023-05-04 09:05:19,645 - fb_one_to_x - [INFO] - [Test, Head_Batch Step 0]	fb_one_to_x
+2023-05-04 09:06:07,591 - fb_one_to_x - [INFO] - [Test, Head_Batch Step 100]	fb_one_to_x
+2023-05-04 09:06:35,660 - fb_one_to_x - [INFO] - [Evaluating Epoch 0 test]: 
+	MRR: Tail : 0.43029, Head : 0.23256, Avg : 0.33142
+	MR: Tail : 149.91, Head : 288.48, Avg : 219.2
+	Hit-1: Tail : 0.33563, Head : 0.14883, Avg : 0.24223
+	Hit-3: Tail : 0.47068, Head : 0.25515, Avg : 0.36292
+	Hit-10: Tail : 0.61952, Head : 0.40096, Avg : 0.51024

log/icews14_128

@@ -0,0 +1 @@
+2023-05-13 03:52:44,141 - icews14_128 - [INFO] - {'dataset': 'icews14', 'name': 'icews14_128', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': True, 'filtered': False}

log/poofnet.log

@@ -0,0 +1,2 @@
+nohup: ignoring input
+python: can't open file 'run.py': [Errno 2] No such file or directory

log/testrun_227cb2f9

@@ -0,0 +1 @@
+2023-05-17 06:54:57,988 - testrun_227cb2f9 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_227cb2f9', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_30d70322

@@ -0,0 +1 @@
+2023-05-17 06:23:34,181 - testrun_30d70322 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_30d70322', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_3212b281

@@ -0,0 +1 @@
+2023-05-17 06:53:01,668 - testrun_3212b281 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_3212b281', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_3dbc9e89

@@ -0,0 +1 @@
+2023-05-06 08:35:38,753 - testrun_3dbc9e89 - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_3dbc9e89', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False}

log/testrun_43389ddf

@@ -0,0 +1 @@
+2023-05-17 06:38:00,469 - testrun_43389ddf - [INFO] - {'dataset': 'icews14', 'name': 'testrun_43389ddf', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_47ede3b9

@@ -0,0 +1 @@
+2023-05-17 06:13:02,952 - testrun_47ede3b9 - [INFO] - {'dataset': 'FB15k-237', 'name': 'testrun_47ede3b9', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_49495af8

@@ -0,0 +1 @@
+2023-05-06 08:37:18,939 - testrun_49495af8 - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_49495af8', 'gpu': '3', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False}

log/testrun_4f5d8391

@@ -0,0 +1 @@
+2023-05-06 08:35:13,356 - testrun_4f5d8391 - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_4f5d8391', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False}

log/testrun_540f6a03

@@ -0,0 +1 @@
+2023-05-06 08:34:55,992 - testrun_540f6a03 - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_540f6a03', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False}

log/testrun_5a901712

@@ -0,0 +1 @@
+2023-05-17 07:04:56,051 - testrun_5a901712 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_5a901712', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_5cafe61a

@@ -0,0 +1,44 @@
+2023-05-17 06:48:57,396 - testrun_5cafe61a - [INFO] - {'dataset': 'icews14', 'name': 'testrun_5cafe61a', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}
+2023-05-17 06:49:44,802 - concurrent.futures - [ERROR] - exception calling callback for <Future at 0x7efb51b74160 state=finished raised BrokenProcessPool>
+joblib.externals.loky.process_executor._RemoteTraceback: 
+"""
+Traceback (most recent call last):
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/externals/loky/process_executor.py", line 391, in _process_worker
+    call_item = call_queue.get(block=True, timeout=timeout)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/multiprocessing/queues.py", line 116, in get
+    return _ForkingPickler.loads(res)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/torch/storage.py", line 222, in _load_from_bytes
+    return torch.load(io.BytesIO(b))
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/torch/serialization.py", line 713, in load
+    return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/torch/serialization.py", line 930, in _legacy_load
+    result = unpickler.load()
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/torch/serialization.py", line 876, in persistent_load
+    wrap_storage=restore_location(obj, location),
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/torch/serialization.py", line 175, in default_restore_location
+    result = fn(storage, location)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/torch/serialization.py", line 155, in _cuda_deserialize
+    return torch._UntypedStorage(obj.nbytes(), device=torch.device(location))
+RuntimeError: CUDA out of memory. Tried to allocate 678.00 MiB (GPU 0; 31.72 GiB total capacity; 0 bytes already allocated; 593.94 MiB free; 0 bytes reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
+"""
+
+The above exception was the direct cause of the following exception:
+
+Traceback (most recent call last):
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/externals/loky/_base.py", line 26, in _invoke_callbacks
+    callback(self)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/parallel.py", line 385, in __call__
+    self.parallel.dispatch_next()
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/parallel.py", line 834, in dispatch_next
+    if not self.dispatch_one_batch(self._original_iterator):
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/parallel.py", line 901, in dispatch_one_batch
+    self._dispatch(tasks)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/parallel.py", line 819, in _dispatch
+    job = self._backend.apply_async(batch, callback=cb)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/_parallel_backends.py", line 556, in apply_async
+    future = self._workers.submit(SafeFunction(func))
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/externals/loky/reusable_executor.py", line 176, in submit
+    return super().submit(fn, *args, **kwargs)
+  File "/opt/conda/envs/kgs2s/lib/python3.8/site-packages/joblib/externals/loky/process_executor.py", line 1129, in submit
+    raise self._flags.broken
+joblib.externals.loky.process_executor.BrokenProcessPool: A task has failed to un-serialize. Please ensure that the arguments of the function are all picklable.

log/testrun_6fd94d59

@@ -0,0 +1 @@
+2023-05-06 08:34:33,652 - testrun_6fd94d59 - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_6fd94d59', 'gpu': '3', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False}

log/testrun_7c096a18

@@ -0,0 +1 @@
+2023-05-17 06:56:35,124 - testrun_7c096a18 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_7c096a18', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_7fb885ee

@@ -0,0 +1 @@
+2023-05-17 07:13:14,777 - testrun_7fb885ee - [INFO] - {'dataset': 'icews14', 'name': 'testrun_7fb885ee', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_8f32040f

@@ -0,0 +1 @@
+2023-05-17 06:59:35,220 - testrun_8f32040f - [INFO] - {'dataset': 'icews14', 'name': 'testrun_8f32040f', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_958ef154

@@ -0,0 +1 @@
+2023-05-17 06:16:45,427 - testrun_958ef154 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_958ef154', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_9acdfb58

@@ -0,0 +1,2 @@
+2023-05-06 08:36:46,668 - testrun_9acdfb58 - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_9acdfb58', 'gpu': '3', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False}
+2023-05-06 08:36:57,409 - testrun_9acdfb58 - [INFO] - [E:0| 0]: Train Loss:0.69813,  Val MRR:0.0, 	testrun_9acdfb58

log/testrun_a051cf32

@@ -0,0 +1 @@
+2023-05-17 06:36:14,606 - testrun_a051cf32 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_a051cf32', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_a06d39d0

@@ -0,0 +1 @@
+2023-05-17 06:13:16,274 - testrun_a06d39d0 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_a06d39d0', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_aca2b734

@@ -0,0 +1 @@
+2023-05-17 06:41:20,654 - testrun_aca2b734 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_aca2b734', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_ad7a0edb

@@ -0,0 +1 @@
+2023-05-17 06:45:54,332 - testrun_ad7a0edb - [INFO] - {'dataset': 'icews14', 'name': 'testrun_ad7a0edb', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_b381870f

@@ -0,0 +1 @@
+2023-05-30 17:54:20,857 - testrun_b381870f - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_b381870f', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0003, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': False}

log/testrun_b396dcde

@@ -0,0 +1,2 @@
+2023-05-30 17:56:25,430 - testrun_b396dcde - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_b396dcde', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0003, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': False}
+2023-05-30 17:57:00,673 - testrun_b396dcde - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_b396dcde', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0003, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': False, 'num_ent': 12554, 'num_rel': 423}

log/testrun_bbf65ab5

@@ -0,0 +1 @@
+2023-05-17 06:21:14,228 - testrun_bbf65ab5 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_bbf65ab5', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_bfaa042b

@@ -0,0 +1 @@
+2023-05-17 06:50:58,251 - testrun_bfaa042b - [INFO] - {'dataset': 'icews14', 'name': 'testrun_bfaa042b', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_c77a8ec3

@@ -0,0 +1 @@
+2023-05-17 06:37:11,288 - testrun_c77a8ec3 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_c77a8ec3', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_cb3528f3

@@ -0,0 +1 @@
+2023-05-17 07:08:13,688 - testrun_cb3528f3 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_cb3528f3', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_cd333c33

@@ -0,0 +1 @@
+2023-05-17 06:25:12,047 - testrun_cd333c33 - [INFO] - {'dataset': 'icews14', 'name': 'testrun_cd333c33', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_d0367b19

@@ -0,0 +1,2 @@
+2023-05-06 08:37:25,129 - testrun_d0367b19 - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_d0367b19', 'gpu': '3', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False}
+2023-05-06 08:37:36,239 - testrun_d0367b19 - [INFO] - [E:0| 0]: Train Loss:0.69813,  Val MRR:0.0, 	testrun_d0367b19

log/testrun_f0394b3c

@@ -0,0 +1 @@
+2023-05-17 06:47:48,537 - testrun_f0394b3c - [INFO] - {'dataset': 'icews14', 'name': 'testrun_f0394b3c', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/testrun_f42f568c

@@ -0,0 +1 @@
+2023-05-30 17:55:52,461 - testrun_f42f568c - [INFO] - {'dataset': 'wikidata12k', 'name': 'testrun_f42f568c', 'gpu': '0', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0003, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': False}

log/testrun_fdb0e82c

@@ -0,0 +1 @@
+2023-05-17 06:39:01,301 - testrun_fdb0e82c - [INFO] - {'dataset': 'icews14', 'name': 'testrun_fdb0e82c', 'gpu': '1', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': True}

log/wikidata12k_0.00003

@@ -0,0 +1,2 @@
+2023-06-04 17:05:45,012 - wikidata12k_0.00003 - [INFO] - {'dataset': 'wikidata12k', 'name': 'wikidata12k_0.00003', 'gpu': '2', 'train_strategy': 'one_to_n', 'opt': 'adam', 'neg_num': 1000, 'batch_size': 128, 'l2': 0.0, 'lr': 0.0001, 'max_epochs': 500, 'num_workers': 0, 'seed': 42, 'restore': False, 'lbl_smooth': 0.1, 'embed_dim': 400, 'ent_vec_dim': 400, 'rel_vec_dim': 400, 'bias': False, 'form': 'plain', 'k_w': 10, 'k_h': 20, 'num_filt': 96, 'ker_sz': 9, 'perm': 1, 'hid_drop': 0.5, 'feat_drop': 0.2, 'inp_drop': 0.2, 'drop_path': 0.0, 'drop': 0.0, 'in_channels': 1, 'out_channels': 32, 'filt_h': 1, 'filt_w': 9, 'image_h': 128, 'image_w': 128, 'patch_size': 8, 'mixer_dim': 256, 'expansion_factor': 4, 'expansion_factor_token': 0.5, 'mixer_depth': 16, 'mixer_dropout': 0.2, 'log_dir': './log/', 'config_dir': './config/', 'test_only': False, 'grid_search': False}
+2023-06-04 17:06:06,702 - wikidata12k_0.00003 - [INFO] - [E:0| 0]: Train Loss:0.69813,  Val MRR:0.0, 	wikidata12k_0.00003

main.py

@@ -3,9 +3,12 @@ import uuid
 import argparse
 import logging
 import logging.config
+import pandas as pd
+import sys
 
 import torch
 import numpy as np
+import time
 
 from collections import defaultdict as ddict
 from pprint import pprint
@@ -17,11 +20,12 @@ from data_loader import TrainDataset, TestDataset
 from utils import get_logger, get_combined_results, set_gpu, prepare_env, set_seed
 
 from models import ComplEx, ConvE, HypER, InteractE, FouriER, TuckER
+import traceback
 
 
 class Main(object):
 
-    def __init__(self, params):
+    def __init__(self, params, logger):
         """
         Constructor of the runner class
         Parameters
@@ -34,11 +38,9 @@ class Main(object):
 
         """
         self.p = params
-        self.logger = get_logger(
-            self.p.name, self.p.log_dir, self.p.config_dir)
+        self.logger = logger
 
         self.logger.info(vars(self.p))
-        pprint(vars(self.p))
 
         if self.p.gpu != '-1' and torch.cuda.is_available():
             self.device = torch.device('cuda')
@@ -76,14 +78,14 @@ class Main(object):
         ent_set, rel_set = OrderedSet(), OrderedSet()
         for split in ['train', 'test', 'valid']:
             for line in open('./data/{}/{}.txt'.format(self.p.dataset, split)):
-                sub, rel, obj = map(str.lower, line.strip().split('\t'))
+                sub, rel, obj, *_ = map(str.lower, line.strip().split('\t'))
                 ent_set.add(sub)
                 rel_set.add(rel)
                 ent_set.add(obj)
         
         self.ent2id = {}
         for line in open('./data/{}/{}'.format(self.p.dataset, "entities.dict")):
-            id, ent = map(str.lower, line.strip().split('\t'))
+            id, ent = map(str.lower, line.replace('\xa0', '').strip().split('\t'))
             self.ent2id[ent] = int(id)
         self.rel2id = {}
         for line in open('./data/{}/{}'.format(self.p.dataset, "relations.dict")):
@@ -108,7 +110,7 @@ class Main(object):
 
         for split in ['train', 'test', 'valid']:
             for line in open('./data/{}/{}.txt'.format(self.p.dataset, split)):
-                sub, rel, obj = map(str.lower, line.strip().split('\t'))
+                sub, rel, obj, *_ = map(str.lower, line.replace('\xa0', '').strip().split('\t'))
                 sub, rel, obj = self.ent2id[sub], self.rel2id[rel], self.ent2id[obj]
                 self.data[split].append((sub, rel, obj))
 
@@ -406,6 +408,13 @@ class Main(object):
             train_iter = iter(
                 self.data_iter['{}_{}'.format(split, mode.split('_')[0])])
 
+            sub_all = []
+            obj_all = []
+            rel_all = []
+            target_score = []
+            target_rank = []
+            obj_pred = []
+            obj_pred_score = []
             for step, batch in enumerate(train_iter):
                 sub, rel, obj, label = self.read_batch(batch, split)
                 pred = self.model.forward(sub, rel, None, 'one_to_n')
@@ -413,9 +422,21 @@ class Main(object):
                 target_pred = pred[b_range, obj]
                 pred = torch.where(label.byte(), torch.zeros_like(pred), pred)
                 pred[b_range, obj] = target_pred
+
+                highest = torch.argsort(pred, dim=1, descending=True)[:,0]
+                highest_score = pred[b_range, highest]
+
                 ranks = 1 + torch.argsort(torch.argsort(pred, dim=1,
                                           descending=True), dim=1, descending=False)[b_range, obj]
 
+                sub_all.extend(sub.cpu().numpy())
+                obj_all.extend(obj.cpu().numpy())
+                rel_all.extend(rel.cpu().numpy())
+                target_score.extend(target_pred.cpu().numpy())
+                target_rank.extend(ranks.cpu().numpy())
+                obj_pred.extend(highest.cpu().numpy())
+                obj_pred_score.extend(highest_score.cpu().numpy())
+
                 ranks = ranks.float()
                 results['count'] = torch.numel(
                     ranks) + results.get('count', 0.0)
@@ -430,7 +451,8 @@ class Main(object):
                 if step % 100 == 0:
                     self.logger.info('[{}, {} Step {}]\t{}'.format(
                         split.title(), mode.title(), step, self.p.name))
-
+        df = pd.DataFrame({"sub":sub_all,"rel":rel_all,"obj":obj_all, "rank": target_rank,"score":target_score, "pred":obj_pred,"pred_score":obj_pred_score})
+        df.to_csv(f"{self.p.name}_result.csv",header=True, index=False)
         return results
 
     def run_epoch(self, epoch):
@@ -456,7 +478,11 @@ class Main(object):
                 batch, 'train')
 
             pred = self.model.forward(sub, rel, neg_ent, self.p.train_strategy)
-            loss = self.model.loss(pred, label, sub_samp)
+            try:
+                loss = self.model.loss(pred, label, sub_samp)
+            except Exception as e:
+                print(pred)
+                raise e
 
             loss.backward()
             self.optimizer.step()
@@ -634,9 +660,10 @@ if __name__ == "__main__":
     set_gpu(args.gpu)
     set_seed(args.seed)
 
-    model = Main(args)
 
     if (args.grid_search):
+        
+        model = Main(args)
         from sklearn.model_selection import GridSearchCV
         from skorch import NeuralNet
 
@@ -685,9 +712,27 @@ if __name__ == "__main__":
             search = grid.fit(inputs, label)
             print("BEST SCORE: ", search.best_score_)
             print("BEST PARAMS: ", search.best_params_)
+    logger = get_logger(
+            args.name, args.log_dir, args.config_dir)
     if (args.test_only):
+        model = Main(args, logger)
         save_path = os.path.join('./torch_saved', args.name)
         model.load_model(save_path)
         model.evaluate('test')
     else:
+        model = Main(args, logger)
         model.fit()
+        # while True:
+        #     try:
+        #         model = Main(args, logger)
+        #         model.fit()
+        #     except Exception as e:
+        #         print(e)
+        #         traceback.print_exc()
+        #         try:
+        #             del model
+        #         except Exception:
+        #             pass
+        #         time.sleep(30)
+        #         continue
+        #     break

models.py

@@ -1,15 +1,16 @@
 import torch
-from torch import nn
+from torch import nn, einsum
 import torch.nn.functional as F
 import numpy as np
 from functools import partial
 from einops.layers.torch import Rearrange, Reduce
+from einops import rearrange, repeat
 from utils import *
 from layers import *
 from timm.data import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
 from timm.models.layers import DropPath, trunc_normal_
 from timm.models.registry import register_model
-from timm.models.layers.helpers import to_2tuple
+from timm.layers.helpers import to_2tuple
 
 
 class ConvE(torch.nn.Module):
@@ -557,6 +558,8 @@ class FouriER(torch.nn.Module):
         z = self.forward_embeddings(y)
         z = self.forward_tokens(z)
         z = z.mean([-2, -1])
+        if np.count_nonzero(np.isnan(z)) > 0:
+            print("ZZZ")
         z = self.norm(z)
         x = self.head(z)
         x = self.hidden_drop(x)
@@ -707,6 +710,363 @@ def basic_blocks(dim, index, layers,
 
     return blocks
 
+def window_partition(x, window_size):
+    """
+    Args:
+        x: (B, H, W, C)
+        window_size (int): window size
+
+    Returns:
+        windows: (num_windows*B, window_size, window_size, C)
+    """
+    B, C, H, W = x.shape
+    x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
+    windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
+    return windows
+
+class WindowAttention(nn.Module):
+    r""" Window based multi-head self attention (W-MSA) module with relative position bias.
+    It supports both of shifted and non-shifted window.
+
+    Args:
+        dim (int): Number of input channels.
+        window_size (tuple[int]): The height and width of the window.
+        num_heads (int): Number of attention heads.
+        qkv_bias (bool, optional):  If True, add a learnable bias to query, key, value. Default: True
+        attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
+        proj_drop (float, optional): Dropout ratio of output. Default: 0.0
+        pretrained_window_size (tuple[int]): The height and width of the window in pre-training.
+    """
+
+    def __init__(self, dim, window_size, num_heads, qkv_bias=True, attn_drop=0., proj_drop=0.,
+                 pretrained_window_size=[0, 0]):
+
+        super().__init__()
+        self.dim = dim
+        self.window_size = window_size  # Wh, Ww
+        self.pretrained_window_size = pretrained_window_size
+        self.num_heads = num_heads
+
+        self.logit_scale = nn.Parameter(torch.log(10 * torch.ones((num_heads, 1, 1))), requires_grad=True)
+
+        # mlp to generate continuous relative position bias
+        self.cpb_mlp = nn.Sequential(nn.Linear(2, 512, bias=True),
+                                     nn.ReLU(inplace=True),
+                                     nn.Linear(512, num_heads, bias=False))
+
+        # get relative_coords_table
+        relative_coords_h = torch.arange(-(self.window_size[0] - 1), self.window_size[0], dtype=torch.float32)
+        relative_coords_w = torch.arange(-(self.window_size[1] - 1), self.window_size[1], dtype=torch.float32)
+        relative_coords_table = torch.stack(
+            torch.meshgrid([relative_coords_h,
+                            relative_coords_w])).permute(1, 2, 0).contiguous().unsqueeze(0)  # 1, 2*Wh-1, 2*Ww-1, 2
+        if pretrained_window_size[0] > 0:
+            relative_coords_table[:, :, :, 0] /= (pretrained_window_size[0] - 1)
+            relative_coords_table[:, :, :, 1] /= (pretrained_window_size[1] - 1)
+        else:
+            relative_coords_table[:, :, :, 0] /= (self.window_size[0] - 1)
+            relative_coords_table[:, :, :, 1] /= (self.window_size[1] - 1)
+        relative_coords_table *= 8  # normalize to -8, 8
+        relative_coords_table = torch.sign(relative_coords_table) * torch.log2(
+            torch.abs(relative_coords_table) + 1.0) / np.log2(8)
+
+        self.register_buffer("relative_coords_table", relative_coords_table)
+
+        # get pair-wise relative position index for each token inside the window
+        coords_h = torch.arange(self.window_size[0])
+        coords_w = torch.arange(self.window_size[1])
+        coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Ww
+        coords_flatten = torch.flatten(coords, 1)  # 2, Wh*Ww
+        relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Ww
+        relative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2
+        relative_coords[:, :, 0] += self.window_size[0] - 1  # shift to start from 0
+        relative_coords[:, :, 1] += self.window_size[1] - 1
+        relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
+        relative_position_index = relative_coords.sum(-1)  # Wh*Ww, Wh*Ww
+        self.register_buffer("relative_position_index", relative_position_index)
+
+        self.qkv = nn.Linear(dim, dim * 3, bias=False)
+        if qkv_bias:
+            self.q_bias = nn.Parameter(torch.zeros(dim))
+            self.v_bias = nn.Parameter(torch.zeros(dim))
+        else:
+            self.q_bias = None
+            self.v_bias = None
+        self.attn_drop = nn.Dropout(attn_drop)
+        self.proj = nn.Linear(dim, dim)
+        self.proj_drop = nn.Dropout(proj_drop)
+        self.softmax = nn.Softmax(dim=-1)
+
+    def forward(self, x, mask=None):
+        """
+        Args:
+            x: input features with shape of (num_windows*B, N, C)
+            mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
+        """
+        B_, N, C = x.shape
+        qkv_bias = None
+        if self.q_bias is not None:
+            qkv_bias = torch.cat((self.q_bias, torch.zeros_like(self.v_bias, requires_grad=False), self.v_bias))
+        qkv = F.linear(input=x, weight=self.qkv.weight, bias=qkv_bias)
+        qkv = qkv.reshape(B_, N, 3, self.num_heads, -1).permute(2, 0, 3, 1, 4)
+        q, k, v = qkv[0], qkv[1], qkv[2]  # make torchscript happy (cannot use tensor as tuple)
+
+        # cosine attention
+        attn = (F.normalize(q, dim=-1) @ F.normalize(k, dim=-1).transpose(-2, -1))
+        logit_scale = torch.clamp(self.logit_scale, max=torch.log(torch.tensor(1. / 0.01)).cuda()).exp()
+        attn = attn * logit_scale
+
+        relative_position_bias_table = self.cpb_mlp(self.relative_coords_table).view(-1, self.num_heads)
+        relative_position_bias = relative_position_bias_table[self.relative_position_index.view(-1)].view(
+            self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1)  # Wh*Ww,Wh*Ww,nH
+        relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww
+        relative_position_bias = 16 * torch.sigmoid(relative_position_bias)
+        attn = attn + relative_position_bias.unsqueeze(0)
+
+        if mask is not None:
+            nW = mask.shape[0]
+            attn = attn.view(B_ // nW, nW, self.num_heads, N, N) + mask.unsqueeze(1).unsqueeze(0)
+            attn = attn.view(-1, self.num_heads, N, N)
+            attn = self.softmax(attn)
+        else:
+            attn = self.softmax(attn)
+
+        attn = self.attn_drop(attn)
+
+        x = (attn @ v).transpose(1, 2).reshape(B_, N, C)
+        x = self.proj(x)
+        x = self.proj_drop(x)
+        return x
+
+    def extra_repr(self) -> str:
+        return f'dim={self.dim}, window_size={self.window_size}, ' \
+               f'pretrained_window_size={self.pretrained_window_size}, num_heads={self.num_heads}'
+
+    def flops(self, N):
+        # calculate flops for 1 window with token length of N
+        flops = 0
+        # qkv = self.qkv(x)
+        flops += N * self.dim * 3 * self.dim
+        # attn = (q @ k.transpose(-2, -1))
+        flops += self.num_heads * N * (self.dim // self.num_heads) * N
+        #  x = (attn @ v)
+        flops += self.num_heads * N * N * (self.dim // self.num_heads)
+        # x = self.proj(x)
+        flops += N * self.dim * self.dim
+        return flops
+    
+def window_reverse(windows, window_size, H, W):
+    """
+    Args:
+        windows: (num_windows*B, window_size, window_size, C)
+        window_size (int): Window size
+        H (int): Height of image
+        W (int): Width of image
+
+    Returns:
+        x: (B, H, W, C)
+    """
+    B = int(windows.shape[0] / (H * W / window_size / window_size))
+    x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
+    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, -1, H, W)
+    return x
+
+def cast_tuple(val, length = 1):
+    return val if isinstance(val, tuple) else ((val,) * length)
+
+# helper classes
+
+class ChanLayerNorm(nn.Module):
+    def __init__(self, dim, eps = 1e-5):
+        super().__init__()
+        self.eps = eps
+        self.g = nn.Parameter(torch.ones(1, dim, 1, 1))
+        self.b = nn.Parameter(torch.zeros(1, dim, 1, 1))
+
+    def forward(self, x):
+        var = torch.var(x, dim = 1, unbiased = False, keepdim = True)
+        mean = torch.mean(x, dim = 1, keepdim = True)
+        return (x - mean) / (var + self.eps).sqrt() * self.g + self.b
+
+class OverlappingPatchEmbed(nn.Module):
+    def __init__(self, dim_in, dim_out, stride = 2):
+        super().__init__()
+        kernel_size = stride * 2 - 1
+        padding = kernel_size // 2
+        self.conv = nn.Conv2d(dim_in, dim_out, kernel_size, stride = stride, padding = padding)
+
+    def forward(self, x):
+        return self.conv(x)
+
+class PEG(nn.Module):
+    def __init__(self, dim, kernel_size = 3):
+        super().__init__()
+        self.proj = nn.Conv2d(dim, dim, kernel_size = kernel_size, padding = kernel_size // 2, groups = dim, stride = 1)
+
+    def forward(self, x):
+        return self.proj(x) + x
+
+# feedforward
+
+class FeedForwardDSSA(nn.Module):
+    def __init__(self, dim, mult = 4, dropout = 0.):
+        super().__init__()
+        inner_dim = int(dim * mult)
+        self.net = nn.Sequential(
+            ChanLayerNorm(dim),
+            nn.Conv2d(dim, inner_dim, 1),
+            nn.GELU(),
+            nn.Dropout(dropout),
+            nn.Conv2d(inner_dim, dim, 1),
+            nn.Dropout(dropout)
+        )
+    def forward(self, x):
+        return self.net(x)
+
+# attention
+
+class DSSA(nn.Module):
+    def __init__(
+        self,
+        dim,
+        heads = 8,
+        dim_head = 32,
+        dropout = 0.,
+        window_size = 7
+    ):
+        super().__init__()
+        self.heads = heads
+        self.scale = dim_head ** -0.5
+        self.window_size = window_size
+        inner_dim = dim_head * heads
+
+        self.norm = ChanLayerNorm(dim)
+
+        self.attend = nn.Sequential(
+            nn.Softmax(dim = -1),
+            nn.Dropout(dropout)
+        )
+
+        self.to_qkv = nn.Conv1d(dim, inner_dim * 3, 1, bias = False)
+
+        # window tokens
+
+        self.window_tokens = nn.Parameter(torch.randn(dim))
+
+        # prenorm and non-linearity for window tokens
+        # then projection to queries and keys for window tokens
+
+        self.window_tokens_to_qk = nn.Sequential(
+            nn.LayerNorm(dim_head),
+            nn.GELU(),
+            Rearrange('b h n c -> b (h c) n'),
+            nn.Conv1d(inner_dim, inner_dim * 2, 1),
+            Rearrange('b (h c) n -> b h n c', h = heads),
+        )
+
+        # window attention
+
+        self.window_attend = nn.Sequential(
+            nn.Softmax(dim = -1),
+            nn.Dropout(dropout)
+        )
+
+        self.to_out = nn.Sequential(
+            nn.Conv2d(inner_dim, dim, 1),
+            nn.Dropout(dropout)
+        )
+
+    def forward(self, x):
+        """
+        einstein notation
+
+        b - batch
+        c - channels
+        w1 - window size (height)
+        w2 - also window size (width)
+        i - sequence dimension (source)
+        j - sequence dimension (target dimension to be reduced)
+        h - heads
+        x - height of feature map divided by window size
+        y - width of feature map divided by window size
+        """
+
+        batch, height, width, heads, wsz = x.shape[0], *x.shape[-2:], self.heads, self.window_size
+        assert (height % wsz) == 0 and (width % wsz) == 0, f'height {height} and width {width} must be divisible by window size {wsz}'
+        num_windows = (height // wsz) * (width // wsz)
+
+        x = self.norm(x)
+
+        # fold in windows for "depthwise" attention - not sure why it is named depthwise when it is just "windowed" attention
+
+        x = rearrange(x, 'b c (h w1) (w w2) -> (b h w) c (w1 w2)', w1 = wsz, w2 = wsz)
+
+        # add windowing tokens
+
+        w = repeat(self.window_tokens, 'c -> b c 1', b = x.shape[0])
+        x = torch.cat((w, x), dim = -1)
+
+        # project for queries, keys, value
+
+        q, k, v = self.to_qkv(x).chunk(3, dim = 1)
+
+        # split out heads
+
+        q, k, v = map(lambda t: rearrange(t, 'b (h d) ... -> b h (...) d', h = heads), (q, k, v))
+
+        # scale
+
+        q = q * self.scale
+
+        # similarity
+
+        dots = einsum('b h i d, b h j d -> b h i j', q, k)
+
+        # attention
+
+        attn = self.attend(dots)
+
+        # aggregate values
+
+        out = torch.matmul(attn, v)
+
+        # split out windowed tokens
+
+        window_tokens, windowed_fmaps = out[:, :, 0], out[:, :, 1:]
+
+        # early return if there is only 1 window
+
+        if num_windows == 1:
+            fmap = rearrange(windowed_fmaps, '(b x y) h (w1 w2) d -> b (h d) (x w1) (y w2)', x = height // wsz, y = width // wsz, w1 = wsz, w2 = wsz)
+            return self.to_out(fmap)
+
+        # carry out the pointwise attention, the main novelty in the paper
+
+        window_tokens = rearrange(window_tokens, '(b x y) h d -> b h (x y) d', x = height // wsz, y = width // wsz)
+        windowed_fmaps = rearrange(windowed_fmaps, '(b x y) h n d -> b h (x y) n d', x = height // wsz, y = width // wsz)
+
+        # windowed queries and keys (preceded by prenorm activation)
+
+        w_q, w_k = self.window_tokens_to_qk(window_tokens).chunk(2, dim = -1)
+
+        # scale
+
+        w_q = w_q * self.scale
+
+        # similarities
+
+        w_dots = einsum('b h i d, b h j d -> b h i j', w_q, w_k)
+
+        w_attn = self.window_attend(w_dots)
+
+        # aggregate the feature maps from the "depthwise" attention step (the most interesting part of the paper, one i haven't seen before)
+
+        aggregated_windowed_fmap = einsum('b h i j, b h j w d -> b h i w d', w_attn, windowed_fmaps)
+
+        # fold back the windows and then combine heads for aggregation
+
+        fmap = rearrange(aggregated_windowed_fmap, 'b h (x y) (w1 w2) d -> b (h d) (x w1) (y w2)', x = height // wsz, y = width // wsz, w1 = wsz, w2 = wsz)
+        return self.to_out(fmap)
 
 class PoolFormerBlock(nn.Module):
     """
@@ -731,7 +1091,15 @@ class PoolFormerBlock(nn.Module):
 
         self.norm1 = norm_layer(dim)
         #self.token_mixer = Pooling(pool_size=pool_size)
-        self.token_mixer = FNetBlock()
+        # self.token_mixer = FNetBlock()
+        self.window_size = 4
+        self.attn_heads = 4
+        self.attn_mask = None
+        # self.token_mixer = WindowAttention(dim=dim, window_size=to_2tuple(self.window_size), num_heads=4)
+        self.token_mixer = nn.ModuleList([
+            DSSA(dim, heads=self.attn_heads, window_size=self.window_size),
+            FeedForwardDSSA(dim)
+        ])
         self.norm2 = norm_layer(dim)
         mlp_hidden_dim = int(dim * mlp_ratio)
         self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, 
@@ -748,16 +1116,26 @@ class PoolFormerBlock(nn.Module):
                 layer_scale_init_value * torch.ones((dim)), requires_grad=True)
 
     def forward(self, x):
+        B, C, H, W = x.shape
+        # x_windows = window_partition(x, self.window_size)
+        # x_windows = x_windows.view(-1, self.window_size * self.window_size, C)
+        # attn_windows = self.token_mixer(x_windows, mask=self.attn_mask)
+        # attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
+        # x_attn = window_reverse(attn_windows, self.window_size, H, W)
+        x_attn = self.token_mixer(x)
         if self.use_layer_scale:
             x = x + self.drop_path(
                 self.layer_scale_1.unsqueeze(-1).unsqueeze(-1)
-                * self.token_mixer(self.norm1(x)))
+                * x_attn)
             x = x + self.drop_path(
                 self.layer_scale_2.unsqueeze(-1).unsqueeze(-1)
                 * self.mlp(self.norm2(x)))
         else:
-            x = x + self.drop_path(self.token_mixer(self.norm1(x)))
+            x = x + self.drop_path(x_attn)
             x = x + self.drop_path(self.mlp(self.norm2(x)))
+
+        if np.count_nonzero(np.isnan(x)) > 0:
+            print("PFBlock")
         return x
 class PatchEmbed(nn.Module):
     """
@@ -843,7 +1221,7 @@ class LayerNormChannel(nn.Module):
             + self.bias.unsqueeze(-1).unsqueeze(-1)
         return x
 
-class FeedForward(nn.Module):
+class FeedForwardFNet(nn.Module):
     def __init__(self, dim, hidden_dim, dropout = 0.):
         super().__init__()
         self.net = nn.Sequential(
@@ -879,7 +1257,7 @@ class FNet(nn.Module):
         for _ in range(depth):
             self.layers.append(nn.ModuleList([
                 PreNorm(dim, FNetBlock()),
-                PreNorm(dim, FeedForward(dim, mlp_dim, dropout = dropout))
+                PreNorm(dim, FeedForwardFNet(dim, mlp_dim, dropout = dropout))
             ]))
     def forward(self, x):
         for attn, ff in self.layers:

requirements.txt

@@ -2,3 +2,5 @@ torch==1.12.1+cu116
 ordered-set==4.1.0
 numpy==1.21.5
 einops==0.4.1
+pandas
+timm==0.9.16

run.sh

@@ -24,3 +24,20 @@ PID: 4503
 test: testrun_d542676f
 ---
 nohup python main.py --gpu 3 --data WN18RR --drop 0.0 --drop_path 0.0 >run_log/fnet-wn.log 2>&1 &
+---
+nohup python main.py --name ice0003 --lr 0.0003 --data icews14 --gpu 1 >run_log/ice0003.log 2>&1 &
+PID: 3076
+tail -f -n 200 run_log/ice0003.log
+---
+nohup python main.py --name ice0003_2 --lr 0.00003 --data icews14 --gpu 3 >run_log/ice0003_2.log 2>&1 &
+PID: 3390
+tail -f -n 200 run_log/ice0003_2.log
+---
+nohup python main.py --name ice00001 --lr 0.00001 --data icews14 --gpu 2 >run_log/ice00001.log 2>&1 &
+PID:
+
+___
+nohup python main.py --name ice001 --lr 0.001 --data icews14 --gpu 3 >run_log/0.001.log 2>&1 &
+___
+nohup python main.py --name iceboth --data icews14_both --gpu 0 >run_log/iceboth.log 2>&1 &
+PID: 21984

visualization_util.py

@@ -0,0 +1,74 @@
+import argparse
+import re
+import os
+import matplotlib.pyplot as plt
+import numpy as np
+from datetime import datetime
+
+def extract_learning_curves(args):
+    paths = args.log_path.split(',')
+    if len(paths) == 1 and os.path.isdir(paths[0]):
+        paths = [os.path.join(paths[0], f) for f in os.listdir(paths[0]) if os.path.isfile(os.path.join(paths[0], f))]
+    learning_curves = {}
+    print(paths)
+    for path in paths:
+        print(path)
+        learning_curve = []
+        lines = open(path, 'r').readlines()
+        last_epoch = -1
+        stacked_epoch = -1
+        max_epoch = -1
+        for line in lines:
+            matched = re.match(r'[0-9\- :,]*\[INFO\] - \[Epoch ([0-9]+)\].*Valid MRR: ([0-9\.]+).*', line)
+            # matched = re.match(r'\tMRR: Tail : [0-9\.]+, Head : [0-9\.]+, Avg : ([0-9\.]+)', line)
+            if matched:
+                this_epoch = int(matched.group(1))
+                if (this_epoch > max_epoch):
+                    learning_curve.append(float(matched.group(2)))
+                    max_epoch = this_epoch
+                    stacked_epoch = this_epoch
+                elif (this_epoch < max_epoch and this_epoch > last_epoch):
+                    last_epoch = this_epoch
+                    max_epoch = stacked_epoch + 1 + this_epoch
+                    learning_curve.append(float(matched.group(2)))
+                if max_epoch >= args.num_epochs:
+                    break
+            # if matched:
+            #     max_epoch += 1
+            #     learning_curve.append(float(matched.group(1)))
+            #     if max_epoch >= args.num_epochs:
+            #         break
+        while len(learning_curve) < args.num_epochs:
+            learning_curve.append(learning_curve[-1])
+        learning_curves[os.path.basename(path)] = learning_curve
+    return learning_curves
+
+def draw_learning_curves(args, learning_curves):
+    for name in learning_curves.keys():
+        epochs = np.arange(len(learning_curves[name]))
+        matched = re.match(r'(.*)\..*', name)
+        if matched:
+            label = matched.group(1)
+        else:
+            label = name
+        plt.plot(epochs, learning_curves[name], label = label)
+    plt.xlabel("Epochs")
+    plt.ylabel("Best Valid MRR")
+    plt.legend(title=args.legend_title)
+    plt.savefig(os.path.join(args.out_path, str(round(datetime.utcnow().timestamp() * 1000)) + '.' + args.fig_filetype))
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(
+        description="Parser For Arguments", formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+    parser.add_argument('--task', default = None, type=str)
+    parser.add_argument('--log_path', type=str, default=None)
+    parser.add_argument('--out_path', type=str, default=None)
+    parser.add_argument('--num_epochs', type=int, default=200)
+    parser.add_argument('--legend_title', type=str, default="Learning rate")
+    parser.add_argument('--fig_filetype', type=str, default="svg")
+
+    args = parser.parse_args()
+
+    if (args.task == 'learning_curve'):
+        draw_learning_curves(args, extract_learning_curves(args))