wukevin commited on
Commit
b2bcfac
1 Parent(s): c30e4e6

Add missing files

Browse files
classifier_training.log DELETED
@@ -1,57 +0,0 @@
1
- Git commit: 01fe7efc56319afc04ab13136d485a680a1b8a82
2
- PyTorch CUDA version: 10.2
3
- Parameter pretrained: ../bert_reduced_intermediate_pe_50_epochs_VDJdb_PIRD
4
- Parameter data: ['PIRD']
5
- Parameter segment: TRB
6
- Parameter outdir: /home/wukevin/projects/tcr/tcr_models/temp3
7
- Parameter downsample: 1.0
8
- Parameter monitor: auprc
9
- Parameter bs: 128
10
- Parameter lr: 5e-05
11
- Parameter epochs: 25
12
- Parameter warmup: 0.1
13
- PIRD data 0.1655 data labelled with antigen sequence
14
- PIRD: Removing 95 entires with non amino acid residues
15
- Entries with antigen sequence: 8429/51044
16
- Unique antigen sequences: 73
17
- PIRD data TRA/TRB instances: Counter({'TRB': 46428, 'TRA': 4011, 'TRA-TRB': 605})
18
- No alphabet given, assuming alphabet of: ['TAAQAAVVRFQEAAN', 'ADTLQSIGATTVASN', 'LLWNGPMAV', 'RPRGEVRFL', 'ATDALMTGY', 'CINGVCWTV', 'HSKKKCDEL', 'KAFSPEVIPMF', 'KRWIILGLNK', 'KRWIIMGLNK', 'TPQDLNTML', 'EIYKRWII', 'GLNKIVRMY', 'ISPRTL-W', 'QASQEVKNW', 'FLKEKGGL', 'HPKVSSEVHI', 'IIKDYGKQM', 'LPPIVAKEI', 'RFPLTFGWCF', 'RYPLTFGWCF', 'TPGPGVRYPL', 'TQGYFPDWQNY', 'FPRPWLHGL', 'RYPLTFGWCF ', 'RYPLTFGW ', 'RYPLTFGW', 'CVETMCNEY', 'DEEDAIAAY', 'ELKRKMIYM', 'ELRRKMMYM', 'LSEFCRVLCCYVLEE', 'QIKVRVDMV', 'QIKVRVKMV', 'VLEETSVML', 'YILEETSVM', 'ARNLVPMVATVQGQN', 'CPSQEPMSIYVY', 'EDVPSGKLFMHVTLG', 'FPTKDVAL', 'IPSINVHHY', 'NLVPMVATV', 'RPHERNGFTVL', 'TPRVTGGGAM', 'YSEHPTFTSQY', 'VTEHDTLLY', 'YLAMPFATPMEAELARRSLA', 'ALTPVVVTL', 'GLCTLVAML', 'YVLDHLIVV', 'EPLPQGQLTAY', 'GPEPLPQGQLTAY', 'LPEGLPQGQLTAY', 'LPEPLGQGQLTAY', 'LPEPLPQAQLTAY', 'LPEPLPQGGLTAY', 'LPEPLPQGQGTAY', 'LPEPLPQGQLGAY', 'LPEPLPQGQLTAY', 'LPEPLPQGQLTGY', 'RAKFKQLL', 'HPVGEADYFEY', 'FLRGRAYGL', 'EENLLDFVRF', 'AVFDRKSDAK', 'IVTDFSVIK', 'NFIRMVISNPAAT', 'KRGIVEQCCTSISSL', 'KRGIVEQSSTSISSL', 'ENPVVHFFKNIVTPR', 'GILGFVFTL', 'PQPELPYPQPE', 'FWIDLFETIG']
19
- Generated labels of shape (6235,): [ 0 0 0 ... 72 72 72]
20
- Retaining 44/73 examples with sufficient examples
21
- Split train with 4365 examples
22
- Split valid with 935 examples
23
- Split test with 935 examples
24
- Loading BERT classifier with 45 labels
25
- Best AUROC (index=27): 0.9405
26
- Best AUROC (index=9): 0.9882
27
- Best AUROC (index=9): 0.9989
28
- Best AUROC (index=9): 1.0000
29
- Best AUROC (index=9): 1.0000
30
- Best AUROC (index=9): 1.0000
31
- Best AUROC (index=9): 1.0000
32
- Best AUROC (index=9): 1.0000
33
- Best AUROC (index=9): 1.0000
34
- Best AUROC (index=9): 1.0000
35
- Best AUROC (index=9): 1.0000
36
- Best AUROC (index=15): 1.0000
37
- Best AUROC (index=9): 1.0000
38
- Best AUROC (index=9): 1.0000
39
- Best AUROC (index=9): 1.0000
40
- Best AUROC (index=9): 1.0000
41
- Best AUROC (index=9): 1.0000
42
- Best AUROC (index=9): 1.0000
43
- Best AUROC (index=9): 1.0000
44
- Best AUROC (index=9): 1.0000
45
- Best AUROC (index=9): 1.0000
46
- Best AUROC (index=9): 1.0000
47
- Best AUROC (index=9): 1.0000
48
- Best AUROC (index=9): 1.0000
49
- Best AUROC (index=21): 1.0000
50
- Best AUROC (index=9): 1.0000
51
- epoch 25.0000
52
- eval_accuracy 0.7401
53
- eval_auprc 0.3922
54
- eval_auroc 0.7881
55
- eval_loss 1.5623
56
- eval_runtime 1.3134
57
- eval_samples_per_second 711.9190
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
params.json DELETED
@@ -1,14 +0,0 @@
1
- {
2
- "pretrained": "../bert_reduced_intermediate_pe_50_epochs_VDJdb_PIRD",
3
- "data": [
4
- "PIRD"
5
- ],
6
- "segment": "TRB",
7
- "outdir": "/home/wukevin/projects/tcr/tcr_models/temp3",
8
- "downsample": 1.0,
9
- "monitor": "auprc",
10
- "bs": 128,
11
- "lr": 5e-05,
12
- "epochs": 25,
13
- "warmup": 0.1
14
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "?", "sep_token": "|", "pad_token": "$", "cls_token": "*", "mask_token": "."}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"do_lower_case": false, "do_basic_tokenize": true, "never_split": null, "unk_token": "?", "sep_token": "|", "pad_token": "$", "cls_token": "*", "mask_token": ".", "tokenize_chinese_chars": false, "strip_accents": null, "model_max_len": 45, "padding_side": "right"}