Skip to content

Commit

Permalink
Wrapping up finetuning updates
Browse files Browse the repository at this point in the history
  • Loading branch information
@WenkelF
WenkelF committed Oct 31, 2024
1 parent 75df01a commit 8436111
Show file tree
Hide file tree
Showing 20 changed files with 23,276 additions and 75 deletions.
13,131 changes: 13,131 additions & 0 deletions expts/data/finetuning_example-cls/raw.csv
View file

Large diffs are not rendered by default.

9,192 changes: 9,192 additions & 0 deletions expts/data/finetuning_example-cls/split.csv
View file

Large diffs are not rendered by default.

161 changes: 161 additions & 0 deletions expts/data/finetuning_example-reg/raw.csv
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,161 @@
,smiles,target
0,CCCS(=O)(=O)Nc1ccc(F)c(C(=O)c2c[nH]c3ncc(-c4ccc(Cl)cc4)cc23)c1F,-1.59345982
1,CCCc1cc(N2CCc3c(nc(C4CC4)n3C)C2)n2ncnc2n1,1.677187053
2,COc1ccccc1C(=O)Nc1ccc2cc[nH]c2c1,0.505149978
3,Cn1cc(Nc2nc(N)nc(-c3cccc(-n4ccc5cc(C6CC6)cc(F)c5c4=O)c3CO)n2)cn1,0.850401148
4,C=CC(=O)N(C)CCOc1c(N)ncnc1-c1cc(F)cc(NC(=O)c2ccc(C3CC3)cc2F)c1C,0.838502776
5,C=CC(=O)N1CCC(CNc2ncnc(N)c2-c2ccc(Oc3ccccc3)cc2)CC1,0.752995432
6,C=CC(=O)N1CCC[C@@H](n2nc(-c3ccc(Oc4ccccc4)cc3)c3c(N)ncnc32)C1,0.62603
7,C=CC(=O)N1CCC[C@H](n2c(=O)n(-c3ccc(Oc4ccccc4)cc3)c3c(N)ncnc32)C1,0.983851719
8,Cc1cccc(/C=N/Nc2cc(N3CCOCC3)nc(OCCc3ccccn3)n2)c1,-0.548213564
9,C=CC(=O)Nc1cccc(Nc2nc(Nc3ccc(Oc4ccnc(C(=O)NC)c4)cc3)ncc2F)c1,-0.403402904
10,C=CC(=O)Nc1cccc(Oc2nc(Nc3ccc(N4CCN(C)CC4)c(F)c3)nc3[nH]ccc23)c1,0.039414119
11,C=CC(=O)Nc1cccc(Oc2nc(Nc3ccc(N4CCN(C)CC4)cc3)nc3ccoc23)c1,0.146128036
12,C=CC(=O)Nc1cccc(Oc2nc(Nc3ccc(N4CCN(C)CC4)cc3)nc3ccsc23)c1,-0.049635146
13,CNC(=O)C1(Cc2ccc(-c3cccnc3)cc2)CCN(Cc2cccc(F)c2)C1,0.829753919
14,COc1nn(C)cc1C(=O)Nc1cccc(-c2cnc3n2CCC3)c1,0.866877814
15,CNC(=O)C1(Cc2ccc(-c3ccncc3)cc2)CCN(Cc2cccc(F)c2)C1,1.10503305
16,COc1nn(C)cc1C(=O)Nc1cccc2cnccc12,1.504729052
17,O=C(Nc1ccc(CN2CCCCC2)cc1)c1ccnc2[nH]cnc12,1.767526899
18,CCN(C/C=C\c1ccc(C2CCCCC2)c(Cl)c1)C1CCCCC1,-0.853871964
19,CC#CC(=O)N1CC[C@@H](n2c(=O)n(-c3ccc(Oc4ccccc4)cc3)c3c(N)ncnc32)C1,0.935859798
20,CC#CC(=O)N[C@H]1CCCN(c2c(F)cc(C(N)=O)c3[nH]c(C)c(C)c23)C1,0.163856803
21,O=C(Nc1ccc2ccccc2n1)c1ccc(N2CCOC2=O)cc1,0.449015316
22,O=C(Nc1cccc(-c2nncn2C2CC2)c1)c1cc(-n2cnc(C3CC3)c2)ccn1,0.262925469
23,O=C(Nc1cccc(N2CCNC2=O)c1)C1NCC12CCCC2,1.718202645
24,Fc1ccc(-c2ccc3c(c2)[nH]c2ccncc23)cn1,0.653405491
25,CC(=O)N1CCN(c2c(Cl)cccc2NC(=O)COc2ccccc2Cl)CC1,0.256958153
26,CC(=O)N1CCN(c2nc(C(F)(F)F)nc3sc(C)c(C)c23)CC1,0.13481437
27,CNc1cc(Nc2cccn(-c3ccccn3)c2=O)nn2c(C(=O)N[C@@H]3C[C@@H]3F)cnc12,0.991226076
28,CCN1CCN(S(=O)(=O)Cc2ccc(Cl)c(Cl)c2)CC1,1.249124949
29,CNc1nc(C)cc(C(=O)Nc2ccc3[nH]ncc3c2)n1,1.493625323
30,Fc1ccccc1-c1c[nH]nc1C1CCCN1Cc1ccc2ncccc2c1,0.545801757
31,O=C(Nc1cnccc1-c1ccc(Cl)cc1)c1ccnc(NC(=O)C2CC2)c1,0.369957607
32,O=C(Nc1nc2cccc(-c3ccc(CN4CCS(=O)(=O)CC4)cc3)n2n1)C1CC1,1.681349797
33,C[C@@H]1CCN(C(=O)CC#N)C[C@@H]1N(C)c1ncnc2[nH]ccc12,1.991815076
34,C[C@@H]1c2nnn(-c3ncc(F)cn3)c2CCN1C(=O)c1cccc(C(F)(F)F)c1Cl,0.969089603
35,CC(=O)Nc1ccc(C(=O)N2CCCCC2c2nc(N)ncc2-c2ccc(Cl)cc2)cc1,0.596926814
36,CC(=O)Nc1ccc(C(=O)N2CCCCC2c2nc(N)ncc2-c2cccc(Cl)c2)cc1,0.520483533
37,C[C@H]1CN(C2COC2)CCN1c1ccc(Nc2cc(-c3ccnc(N4CCn5c(cc6c5CC(C)(C)C6)C4=O)c3CO)cn(C)c2=O)nc1,1.209139536
38,CC(=O)Nc1ccc(O)cc1,1.887859133
39,Cc1cccnc1Nc1cccc(C2CCCN(CC(=O)Nc3nccs3)C2)n1,0.180699201
40,COCC(=O)N1CCC(Cc2ccccc2-c2cccc(F)c2)(C(=O)NC(C)C)CC1,0.596597096
41,Cc1[nH]nc2c1C1(CCCCC1)CC(=O)N2,1.323066376
42,CC(=O)Nc1ncc(C(=O)O)s1,1.81935965
43,N#Cc1cc(F)c(NS(=O)(=O)c2c[nH]c3cc(Cl)ccc23)cc1F,-0.026872146
44,Cc1c(C(=O)NCCCN(C)C)sc2ncnc(Nc3ccc(F)cc3OC(C)C)c12,0.741821047
45,Cc1c(Cl)ccc2cc3n(c12)[C@@H](C)CNC3=O,0.328379603
46,Cc1cn2nc(-c3cc(=O)n4cc(N5CCNC6(CC6)C5)ccc4n3)cc(C)c2n1,0.985246791
47,CCOc1cc(CC(=O)N[C@@H](CC(C)C)c2ccccc2N2CCCCC2)ccc1C(=O)O,0.106870544
48,COCCNC(=O)c1ccnc(C2CCNCC2)c1,2.0
49,Cc1c[nH]c(=O)n1-c1ccc(C(=O)Nc2ccc3ccccc3n2)cc1,0.639386869
50,Cc1cnc(C(=O)NCCc2ccc(S(=O)(=O)NC(=O)NC3CCCCC3)cc2)cn1,0.01745073
51,Cc1c[nH]c2nccc(Oc3c(F)cc(Nc4cc(Cl)nc(N)n4)cc3F)c12,-1.22184875
52,CC(C)(C)C(=O)N1CCC(Cc2ccc(-c3cccs3)cc2)(C(=O)N2CCCC2)CC1,-0.1837587
53,O=C(c1ccc(Oc2ccccc2)cc1Cl)c1c[nH]c2ncnc(N[C@@H]3CC[C@@H](CO)OC3)c12,-0.467245621
54,CC(C)(C)c1ccc(-c2nc3n(c(=O)c2C#N)CCS3)cc1,0.731266349
55,CC(C)(C)c1ccc(C(O)CN2CCC(O)(c3ccc4c(c3)OCO4)CC2)cc1,0.779018972
56,O=C1CCCC[C@@H]2[C@H](C[C@@H](Cc3ccccc3F)N2C(=O)c2cccc3ncccc23)N1,1.277998644
57,CCc1c[nH]c2ncnc(N3CCC(CN4CCN(C)CC4)CC3)c12,1.728012707
58,CC(C)(Oc1ccc(-c2cnc(N)c(-c3ccc(Cl)cc3)c2)cc1)C(=O)O,-1.158015195
59,NC(=O)c1cnc(N2CCc3[nH]nc(C(F)(F)F)c3C2)c(Cl)c1,0.986009932
60,NC1CC(NC(=O)c2ccc(-c3cn[nH]c3)cn2)C12CCC2,1.709702344
61,NC1CCC(C(=O)N2CCC(c3c[nH]c4ncccc34)CC2)C1,1.917395215
62,NC1CCCC(C(=O)N2CCC(c3c[nH]c4ncccc34)CC2)C1,1.752816431
63,NC1CCCC(C(=O)Nc2ccc3[nH]ncc3c2)C1,2.0
64,NC1CCCC(C(=O)Nc2cccc(N3CCNC3=O)c2)C1,1.763113391
65,NC1CCCC1C(=O)N1CCC(c2c[nH]c3ncccc23)CC1,1.71701274
66,O=C1NCCN(C(=O)c2ccc3nccn3c2)C1c1ccccc1C(F)(F)F,1.851001366
67,O=C1NCCN(C(=O)c2ccncc2)C1c1ccccc1Cl,1.876996793
68,NCC1CCCC1NC(=O)c1cc(N2CCNC2=O)ccc1F,1.87495702
69,O=C1NCCSc2c1sc1ccc(O)cc21,0.891593204
70,NCCN1CCN(C/C=C/C(=O)N2CCC[C@@H](n3nc(-c4ccc(Oc5ccccc5)cc4)c4c(N)ncnc43)C2)CC1,0.699837726
71,CCc1nc(C)cn2nc(-c3cc(=O)n4cc(C5CCN(C)CC5)cc(C)c4n3)cc12,0.868232868
72,O=S(=O)(c1cccc2cnccc12)N1CCCNCC1,1.892077899
73,Nc1c(F)ccc2cnc(-n3ccc4ccncc43)cc12,0.713910354
74,CCc1nc2c(C)cc(N3CCN(CC(=O)N4CC(O)C4)CC3)cn2c1N(C)c1nc(-c2ccc(F)cc2)c(C#N)s1,-0.204728421
75,Cc1nc(Nc2nccs2)cc(C2CN(c3ncccn3)C2)n1,0.444669231
76,CC(C)NC(=O)COc1cccc(-c2nc(Nc3ccc4[nH]ncc4c3)c3ccccc3n2)c1,-1.384078213
77,Cc1cc(F)ccc1C1C(=O)NCCN1C(=O)c1ccc2nccn2c1,1.78096503
78,Cc1cc(N2CCCC2c2cc(CCC(=O)O)cc(C)n2)ncn1,2.0
79,CC(C)[C@H](CO)Nc1nc(Nc2cc(N)cc(Cl)c2)c2ncn(C(C)C)c2n1,-0.061980903
80,CCn1c(-c2nonc2N)nc2cnc(Oc3cccc(NC(=O)c4ccc(OCCN5CCOCC5)cc4)c3)cc21,-0.449771647
81,Nc1ncc(-c2cccc(C(F)(F)F)c2)c(C2CCCCN2C(=O)c2ccccc2)n1,-0.04431225
82,CCn1c(=O)oc2cc(NC(=O)c3ccc(C(C)(C)C)cc3)ccc21,0.21005085
83,Nc1ncc(C(=O)NC2CN(C(=O)C3CC3)C2)c2ccc(-c3cccc(F)c3)nc12,0.623352682
84,CCn1c(CO)nn(-c2cc(O[C@@H](C)C(F)(F)F)c(C(=O)Nc3c(F)cccc3Cl)cc2F)c1=O,0.975753389
85,Nc1ncnc2c1c(-c1ccc(Oc3ccccc3)cc1)nn2[C@@H]1CCCNC1,0.956648579
86,Nc1ncnc2c1c(-c1cnc3[nH]ccc3c1)nn2C1CCCC1,0.862608364
87,CCn1cc(C(=O)O)c(=O)c2c(N)c(F)c(NC3CCCCC3)cc21,0.427486109
88,Cc1nccc(-c2cn(Cc3ccccc3)c3cnccc23)n1,1.122707254
89,Cc1cc(Nc2cnccn2)cc(C2CCCN(CC(=O)N3CCCC3)C2)n1,1.642246824
90,Cc1ncsc1C(=O)N1CCCCC1c1nc(N)ncc1-c1cccc(C(F)(F)F)c1,0.615318657
91,CN(C(=O)c1cc(N2CCNC2=O)ccc1F)C1CCNC1,1.885728632
92,OC[C@H](Nc1cncc(-c2ccc3[nH]ncc3n2)c1)c1ccccc1,1.017826038
93,CN(C)C(=O)C1(Cc2ccccc2-c2cccc(F)c2)CCN(C(=O)C2CC=CCC2)CC1,-0.104025268
94,CN(C)C(=O)C1(Cc2ccccc2-c2ccccc2)CCN(C(=O)C2CCCO2)CC1,0.902546779
95,CN(C)C(=O)C1(Cc2ccccc2-c2ccccc2)CCN(C(=O)c2cnn(C)c2)CC1,1.041353202
96,O=C(CCNC(=O)c1ccc(OC(F)(F)F)cc1)N[C@@H]1CCCc2ccccc21,0.526080692
97,CC(O)(C#Cc1ccc2c(c1)N(c1nc(N)ncc1Cl)CC2)c1nccs1,-0.122628654
98,COc1ccc(CCCN2CCN(c3cnn(C)c3)C(=O)C2)cc1F,1.343408594
99,CC/C(=C(\c1ccccc1)c1ccc(OCCN(C)C)cc1)c1ccccc1,-0.580044252
100,CC1(C)CC(Oc2ccc(-c3ccc(-c4cn[nH]c4)cc3O)nn2)CC(C)(C)N1,1.199947058
101,Cc1nnc(-c2ccc(N3CCC(Oc4cc(F)ccc4Cl)CC3)nn2)o1,0.161068385
102,Cc1nnc(-c2ccc(N3CCC(Oc4ccccc4C(F)(F)F)CC3)nn2)s1,-0.906578315
103,c1ccc(-c2ccc(CN3CCCCCCC3)cc2)cc1,-0.614393726
104,COc1ccc(CNC(=O)c2sc3nc(C)cc(C)c3c2N)cc1,-0.356547324
105,Cc1nnc(CN(C)CC(C)Oc2ccc(Cl)c(Cl)c2)n1C,0.926290987
106,CC1(C)Cc2cc(NC(=O)c3cnn4cccnc34)c(OCC3CC3)nc2O1,-0.199970641
107,O=C(CN1CCCC(c2cccc(Cc3cccc(F)c3)n2)C1)N1CCCC1,1.062130535
108,CC1(CNC(=O)c2cncc(C3CCNCC3)n2)CCCO1,1.988550039
109,Cc1noc(C(C)C)c1C(=O)N1CC(C)OC(c2ccccc2)C1,1.180469962
110,CC1CN(C(=O)c2cccnc2N2CCOCC2)CC(c2ccccc2)O1,1.573185017
111,CCC(=O)N1CCN(c2ccc(Cl)cc2NC(=O)COc2ccccc2)CC1,0.250420002
112,COc1ccc(S(=O)(=O)N2CCC(N3CCC(C)CC3)CC2)cc1,1.729799023
113,COc1ccc([C@H]2CN(C(C)=O)[C@@H]3CCCN(Cc4cccc(F)c4)[C@H]23)cc1,0.595385981
114,Cc1oc2ccccc2c1CNc1nnc(-c2ccncc2)o1,0.717254313
115,c1sc(NCC2CCCO2)nc1C12CC3CC(CC(C3)C1)C2,-1.096910013
116,CCC1=C(C)CN(C(=O)NCCc2ccc(S(=O)(=O)NC(=O)N[C@H]3CC[C@H](C)CC3)cc2)C1=O,-1.180456064
117,CN(Cc1ccccc1)C1(C(=O)N2CCNC(=O)CC2)Cc2ccccc2C1,0.904931827
118,CCCCNC(=O)NS(=O)(=O)c1ccc(C)cc1,0.439332694
119,O=C(NC1CCNCC1)c1ccc2[nH]ncc2c1,1.842696589
120,Cn1c(C2CC2)nc2c1CCN(c1ncnc3ccsc13)C2,1.237065953
121,Cc1ccc(OCC2(O)CCN(CC3(O)CCN(c4ccccc4C)CC3)CC2)cc1,-0.040481623
122,CCCNC(=O)NS(=O)(=O)c1ccc(Cl)cc1,0.969835093
123,Cc1ccc(Oc2ccc(Cl)cc2NC(=O)CN(C)CC(=O)N(C)C)cc1,-0.053056729
124,COc1ccccc1-c1cc(NC(=O)c2cccc(N3CCNC3=O)c2)[nH]n1,-0.397940009
125,CN1C(N)=N[C@](C)(c2cc(NC(=O)c3ccc(F)cn3)ccc2F)CS1(=O)=O,1.560468571
126,C=CC(=O)N1C[C@H](Nc2ncnc3[nH]ccc23)CC[C@@H]1C,1.828227965
127,CC#CC(=O)N1CC[C@@H](n2cc(-c3ccc(Oc4c(F)cccc4F)cc3)c3c(N)n[nH]c(=O)c32)C1,-0.073143291
128,CC(=O)NCCNc1cc(Cl)nn2c(-c3cccc(S(=O)(=O)N(C)C)c3)c(C)nc12,0.681693392
129,CC(C)(O)CCn1cc2cc(NC(=O)c3cccc(C(F)(F)F)n3)c(C(C)(C)O)cc2n1,1.015611205
130,CC1CC(N)CCN1C(=O)c1cc(N2CCNC2=O)c(F)cc1F,1.833504094
131,CCOc1cc2nn(CCC(C)(C)O)cc2cc1NC(=O)c1cccc(C(F)F)n1,0.099680641
132,CCS(=O)(=O)N1CC(CC#N)(n2cc(-c3ncnc4[nH]ccc34)cn2)C1,1.378597739
133,CN(C(=O)c1cc(N2CCNC2=O)ccc1F)C1CCC(N)CC1,1.828717885
134,CN1CCN(S(=O)(=O)c2ccc(-c3cnc(N)c(C(=O)Nc4cccnc4)n3)cc2)CC1,1.078638038
135,CNC(=O)C1(Cc2ccc(-c3cccnc3)cc2)CCN(Cc2ccc(F)cc2)C1,1.021602716
136,CNC(=O)C1(Cc2ccc(-c3cccnc3)cc2)CCN(Cc2ccccc2Cl)C1,0.387567779
137,CNC(=O)c1cccc(NC(=O)N2CCC(Oc3ccccc3Cl)CC2)c1,1.068760828
138,COCCc1noc(CN2CC(c3ccccc3)(c3ccccc3)CCC2=O)n1,1.167051359
139,COc1cc2c(cc1OC)CC(=O)N(CCCN(C)C[C@H]1Cc3cc(OC)c(OC)cc31)CC2,1.730984039
140,COc1cc2ncnc(Nc3cccc(O)c3)c2cc1OC,0.994317153
141,COc1ccc(Cl)cc1C(=O)NCCc1ccc(S(=O)(=O)NC(=O)NC2CCCCC2)cc1,-0.91721463
142,COc1ccc(Nc2c(C#N)cnc3cc(OC)c(OC)cc23)cc1Cl,-0.033858267
143,COc1ccc(OCC(=O)Nc2cccc(Cl)c2N2CCN(C(C)=O)CC2)cc1,0.700790221
144,COc1nc2sc(C(=O)NC3CC3)c(N)c2c(C)c1Cl,0.245512668
145,Cc1cc(C)c2c(N)c(C(=O)NCc3ccc(Cl)cc3)sc2n1,-0.370590401
146,Cc1cc(N2CCCC2)nc(C2CCCN(C)C2)n1,2.0
147,Cc1ccc(C(=O)N2CCC(Cc3ccccc3-c3ccccc3)(C(=O)N(C)C)CC2)s1,-0.293282218
148,Cc1ncc(CN2CCC(Nc3ccc4nnnn4n3)C(C)C2)s1,1.652285029
149,Cc1ncsc1C(=O)N1CCCCC1c1nc(N(C)C)ncc1-c1cccc(Cl)c1,-0.083019953
150,Cn1cc(-c2cn3nccc3c(-c3cnn([C@]4(CC#N)C[C@@H](C#N)C4)c3)n2)cn1,1.559583476
151,N#CC[C@H](C1CCCC1)n1cc(-c2ncnc3[nH]ccc23)cn1,0.630986911
152,NC(=O)C1(Cc2ccc(-c3ccncc3)cc2)CCN(C(=O)Cc2cccc(F)c2)CC1,1.006508828
153,O=C(NC12CCCC1NCC2)c1ccc(-c2cn[nH]c2)cc1,2.0
154,O=C(NCCc1ccccc1)c1ccc(NC(=O)N2CCCCc3ccccc32)cc1,0.075911761
155,O=C1CN(c2ccc(Nc3nccc(C(F)(F)F)n3)cn2)CCN1,1.543819805
156,OCC1CCCCN1Cc1ccc(-c2ccccc2)cc1,1.368007805
157,OCC1CCCCN1Cc1ccc(Cl)c(Cl)c1,1.390069186
158,[2H]C([2H])([2H])NC(=O)c1nnc(NC(=O)C2CC2)cc1Nc1cccc(-c2ncn(C)n2)c1OC,1.186566481
159,c1ccc(Oc2cccc(CN(CCN3CCOCC3)Cc3cccnc3)c2)cc1,-0.222573178
110 changes: 110 additions & 0 deletions expts/data/finetuning_example-reg/split.csv
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,110 @@
,train,val,test
0,0,60.0,126.0
1,1,5.0,127.0
2,2,35.0,128.0
3,3,23.0,129.0
4,4,15.0,130.0
5,6,68.0,131.0
6,7,12.0,132.0
7,8,45.0,133.0
8,9,119.0,134.0
9,10,113.0,135.0
10,11,41.0,136.0
11,13,88.0,137.0
12,16,30.0,138.0
13,17,74.0,139.0
14,18,54.0,140.0
15,19,73.0,141.0
16,20,14.0,142.0
17,21,,143.0
18,22,,144.0
19,24,,145.0
20,25,,146.0
21,26,,147.0
22,27,,148.0
23,28,,149.0
24,29,,150.0
25,31,,151.0
26,32,,152.0
27,33,,153.0
28,34,,154.0
29,36,,155.0
30,37,,156.0
31,38,,157.0
32,39,,158.0
33,40,,159.0
34,42,,
35,43,,
36,44,,
37,46,,
38,47,,
39,48,,
40,49,,
41,50,,
42,51,,
43,52,,
44,53,,
45,55,,
46,56,,
47,57,,
48,58,,
49,59,,
50,61,,
51,62,,
52,63,,
53,64,,
54,65,,
55,66,,
56,67,,
57,69,,
58,70,,
59,71,,
60,72,,
61,75,,
62,76,,
63,77,,
64,78,,
65,79,,
66,80,,
67,81,,
68,82,,
69,83,,
70,84,,
71,85,,
72,86,,
73,87,,
74,89,,
75,90,,
76,91,,
77,92,,
78,93,,
79,94,,
80,95,,
81,96,,
82,97,,
83,98,,
84,99,,
85,100,,
86,101,,
87,102,,
88,103,,
89,104,,
90,105,,
91,106,,
92,107,,
93,108,,
94,109,,
95,110,,
96,111,,
97,112,,
98,114,,
99,115,,
100,116,,
101,117,,
102,118,,
103,120,,
104,121,,
105,122,,
106,123,,
107,124,,
108,125,,
105 changes: 105 additions & 0 deletions expts/hydra-configs/finetuning/custom.yaml
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,105 @@
# @package _global_

defaults:
- override /tasks/loss_metrics_datamodule: finetune

constants:
benchmark: custom
task: task
# task_type: reg
# task_name: adme-fang-HPPB-reg
task_type: cls
task_name: CYP2D6_Veith
wandb:
name: finetune_${constants.task_name}
project: finetuning
entity: valencelabs
tags:
- finetuning
- ${constants.task_name}
- ${finetuning.pretrained_model}
seed: 42
max_epochs: 20
raise_train_error: true
model_dropout: 0.

datamodule:
args:
batch_size_training: 256
batch_size_inference: 256
dataloading_from: ram
persistent_workers: true
num_workers: 8

task_specific_args:
finetune:
df: null
# df_path: expts/data/finetuning_example-reg/raw.csv
# splits_path: expts/data/finetuning_example-reg/split.csv
df_path: expts/data/finetuning_example-cls/raw.csv
splits_path: expts/data/finetuning_example-cls/split.csv
smiles_col: smiles
label_cols: target
task_level: graph
epoch_sampling_fraction: 1.0

trainer:
model_checkpoint:
save_top_k: 0
dirpath: ~/project/outgoing/finetuning/${constants.task}/${now:%Y-%m-%d_%H-%M-%S.%f}/
every_n_epochs: 200
save_last: false
trainer:
precision: 32
check_val_every_n_epoch: 1
accumulate_grad_batches: 1

predictor:
optim_kwargs:
lr: 0.000001
torch_scheduler_kwargs:
module_type: WarmUpLinearLR
max_num_epochs: ${constants.max_epochs}
warmup_epochs: 3
verbose: False



# == Fine-tuning config ==

finetuning:
task: finetune
level: graph
pretrained_model: dummy-pretrained-model
finetuning_module: graph_output_nn
sub_module_from_pretrained: graph
new_sub_module: graph
drop_depth: 1
added_depth: 1
new_out_dim: 1024 # TODO: infer automatically
# new_out_dim: 1024 # TODO: infer automatically

keep_modules_after_finetuning_module:
task_heads-pcba_1328:
new_sub_module: finetune
hidden_dims: 512
depth: 1
dropout: 0.
last_activation: none
out_dim: 1

# Optional finetuning head appended to model after finetuning_module
# finetuning_head:
# task: ${constants.task}
# previous_module: task_heads
# incoming_level: graph
# model_type: mlp
# depth: 2
# in_dim: 3707 # TODO: infer automatically
# hidden_dims: 1024
# out_dim: 1
# dropout: 0.
# last_layer_is_readout: true

epoch_unfreeze_all: 0
freeze_always: []
15 changes: 15 additions & 0 deletions expts/hydra-configs/fingerprinting/custom.yaml
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
pretrained_models:
dummy-pretrained-model:
- graph_output_nn-graph:0
- task_heads-zinc:0

datamodule:
df_path: ./expts/data/finetuning_example-reg/raw.csv
benchmark: null
task: null
split_val: 0.0
split_test: 1.0
device: cpu # cpu or cuda
num_workers: 0
fps_cache_dir: ./expts/data/finetuning_example-reg
mol_cache_dir: ${datamodule.fps_cache_dir}
Loading

0 comments on commit 8436111

Please sign in to comment.