add pre-commit

tests/testdata/squad2-v0-greedy_until

-b261e8885c11750ce6911bb11e8693de03d53758297c26fb14cfc1ef508862cb
\ No newline at end of file
+b261e8885c11750ce6911bb11e8693de03d53758297c26fb14cfc1ef508862cb

tests/testdata/squad2-v0-loglikelihood

-287e87cc6878debcc80d9b6df4e2d0a74ed29068e0e0a80906c8441843a17cee
\ No newline at end of file
+287e87cc6878debcc80d9b6df4e2d0a74ed29068e0e0a80906c8441843a17cee

tests/testdata/squad2-v0-res.json

-{"results": {"squad2": {"HasAns_exact": 0.0, "HasAns_f1": 0.0, "NoAns_exact": 0.0, "NoAns_f1": 0.0, "best_exact": 50.07159100480081, "best_f1": 50.07159100480081, "exact": 0.0, "f1": 0.0}}, "versions": {"squad2": 0}}
\ No newline at end of file
+{"results": {"squad2": {"HasAns_exact": 0.0, "HasAns_f1": 0.0, "NoAns_exact": 0.0, "NoAns_f1": 0.0, "best_exact": 50.07159100480081, "best_f1": 50.07159100480081, "exact": 0.0, "f1": 0.0}}, "versions": {"squad2": 0}}

tests/testdata/squad2-v1-greedy_until

-e17e3d85c1d5adaf2d6b4b752c4babc2e0b3a6e144e6de70cb3b2287e85109b8
\ No newline at end of file
+e17e3d85c1d5adaf2d6b4b752c4babc2e0b3a6e144e6de70cb3b2287e85109b8

tests/testdata/squad2-v1-loglikelihood

-f5da6173402b274dc89130755c222c6ca6b2a3bacaaa4e4ab07be9322b7bad65
\ No newline at end of file
+f5da6173402b274dc89130755c222c6ca6b2a3bacaaa4e4ab07be9322b7bad65

tests/testdata/squad2-v1-res.json

-{"results": {"squad2": {"HasAns_exact": 0.0, "HasAns_f1": 0.0, "NoAns_exact": 0.0, "NoAns_f1": 0.0, "best_exact": 50.07159100480081, "best_f1": 50.07159100480081, "exact": 0.0, "f1": 0.0}}, "versions": {"squad2": 1}}
\ No newline at end of file
+{"results": {"squad2": {"HasAns_exact": 0.0, "HasAns_f1": 0.0, "NoAns_exact": 0.0, "NoAns_f1": 0.0, "best_exact": 50.07159100480081, "best_f1": 50.07159100480081, "exact": 0.0, "f1": 0.0}}, "versions": {"squad2": 1}}

tests/testdata/sst-v0-loglikelihood

-d2ebe3a63517d1d481aa1513bebe124c57a0904554a1e95f566979cfe67b1a7f
\ No newline at end of file
+d2ebe3a63517d1d481aa1513bebe124c57a0904554a1e95f566979cfe67b1a7f

tests/testdata/sst-v0-res.json

-{"results": {"sst": {"acc": 0.5172018348623854, "acc_stderr": 0.016931824425903734}}, "versions": {"sst": 0}}
\ No newline at end of file
+{"results": {"sst": {"acc": 0.5172018348623854, "acc_stderr": 0.016931824425903734}}, "versions": {"sst": 0}}

tests/testdata/swag-v0-loglikelihood

-be4fcbad876124c4ba3c71970538a97fec0e36a9cc677c70b6c9243a7bcee0ec
\ No newline at end of file
+be4fcbad876124c4ba3c71970538a97fec0e36a9cc677c70b6c9243a7bcee0ec

tests/testdata/swag-v0-res.json

-{"results": {"swag": {"acc": 0.2482255323402979, "acc_norm": 0.24882535239428172, "acc_norm_stderr": 0.00305666959496067, "acc_stderr": 0.003054201832644171}}, "versions": {"swag": 0}}
\ No newline at end of file
+{"results": {"swag": {"acc": 0.2482255323402979, "acc_norm": 0.24882535239428172, "acc_norm_stderr": 0.00305666959496067, "acc_stderr": 0.003054201832644171}}, "versions": {"swag": 0}}

tests/testdata/triviaqa-v0-loglikelihood

-f8ec05b306b9f6187c0f8117cae441fb85a7a2e4670f4f9a1a3b632b1978421a
\ No newline at end of file
+f8ec05b306b9f6187c0f8117cae441fb85a7a2e4670f4f9a1a3b632b1978421a

tests/testdata/triviaqa-v0-res.json

-{"results": {"triviaqa": {"acc": 0.0, "acc_stderr": 0.0}}, "versions": {"triviaqa": 0}}
\ No newline at end of file
+{"results": {"triviaqa": {"acc": 0.0, "acc_stderr": 0.0}}, "versions": {"triviaqa": 0}}

tests/testdata/truthfulqa_gen-v0-greedy_until

-0d7c56e1aa71ffd8f94bde28f6e8dfdd35f7aaadffa0620bd2a27704253d6c14
\ No newline at end of file
+0d7c56e1aa71ffd8f94bde28f6e8dfdd35f7aaadffa0620bd2a27704253d6c14

tests/testdata/truthfulqa_gen-v0-res.json

-{"results": {"truthfulqa_gen": {"bleu_acc": 0.0, "bleu_acc_stderr": 0.0, "bleu_diff": 0.0, "bleu_diff_stderr": 0.0, "bleu_max": 0.0, "bleu_max_stderr": 0.0, "bleurt_acc": 0.8372093023255814, "bleurt_acc_stderr": 0.012923696051772253, "bleurt_diff": 0.13967358205134603, "bleurt_diff_stderr": 0.00532907098769571, "bleurt_max": -1.4402793981454072, "bleurt_max_stderr": 0.0021884846359458963, "rouge1_acc": 0.0, "rouge1_acc_stderr": 0.0, "rouge1_diff": 0.0, "rouge1_diff_stderr": 0.0, "rouge1_max": 0.0, "rouge1_max_stderr": 0.0, "rouge2_acc": 0.0, "rouge2_acc_stderr": 0.0, "rouge2_diff": 0.0, "rouge2_diff_stderr": 0.0, "rouge2_max": 0.0, "rouge2_max_stderr": 0.0, "rougeL_acc": 0.0, "rougeL_acc_stderr": 0.0, "rougeL_diff": 0.0, "rougeL_diff_stderr": 0.0, "rougeL_max": 0.0, "rougeL_max_stderr": 0.0}}, "versions": {"truthfulqa_gen": 0}}
\ No newline at end of file
+{"results": {"truthfulqa_gen": {"bleu_acc": 0.0, "bleu_acc_stderr": 0.0, "bleu_diff": 0.0, "bleu_diff_stderr": 0.0, "bleu_max": 0.0, "bleu_max_stderr": 0.0, "bleurt_acc": 0.8372093023255814, "bleurt_acc_stderr": 0.012923696051772253, "bleurt_diff": 0.13967358205134603, "bleurt_diff_stderr": 0.00532907098769571, "bleurt_max": -1.4402793981454072, "bleurt_max_stderr": 0.0021884846359458963, "rouge1_acc": 0.0, "rouge1_acc_stderr": 0.0, "rouge1_diff": 0.0, "rouge1_diff_stderr": 0.0, "rouge1_max": 0.0, "rouge1_max_stderr": 0.0, "rouge2_acc": 0.0, "rouge2_acc_stderr": 0.0, "rouge2_diff": 0.0, "rouge2_diff_stderr": 0.0, "rouge2_max": 0.0, "rouge2_max_stderr": 0.0, "rougeL_acc": 0.0, "rougeL_acc_stderr": 0.0, "rougeL_diff": 0.0, "rougeL_diff_stderr": 0.0, "rougeL_max": 0.0, "rougeL_max_stderr": 0.0}}, "versions": {"truthfulqa_gen": 0}}

tests/testdata/truthfulqa_gen-v1-greedy_until

-1a280973bbac2b7ac29dd64dddac474fb4749585f7de893483b4034814466c67
\ No newline at end of file
+1a280973bbac2b7ac29dd64dddac474fb4749585f7de893483b4034814466c67

tests/testdata/truthfulqa_gen-v1-res.json

-{"results": {"truthfulqa_gen": {"bleu_acc": 0.0, "bleu_acc_stderr": 0.0, "bleu_diff": 0.0, "bleu_diff_stderr": 0.0, "bleu_max": 0.0, "bleu_max_stderr": 0.0, "bleurt_acc": 0.835985312117503, "bleurt_acc_stderr": 0.012962704327492454, "bleurt_diff": 0.14077322143090107, "bleurt_diff_stderr": 0.005459888909582694, "bleurt_max": -1.4399358725752065, "bleurt_max_stderr": 0.0022126992369197133, "rouge1_acc": 0.0, "rouge1_acc_stderr": 0.0, "rouge1_diff": 0.0, "rouge1_diff_stderr": 0.0, "rouge1_max": 0.0, "rouge1_max_stderr": 0.0, "rouge2_acc": 0.0, "rouge2_acc_stderr": 0.0, "rouge2_diff": 0.0, "rouge2_diff_stderr": 0.0, "rouge2_max": 0.0, "rouge2_max_stderr": 0.0, "rougeL_acc": 0.0, "rougeL_acc_stderr": 0.0, "rougeL_diff": 0.0, "rougeL_diff_stderr": 0.0, "rougeL_max": 0.0, "rougeL_max_stderr": 0.0}}, "versions": {"truthfulqa_gen": 1}}
\ No newline at end of file
+{"results": {"truthfulqa_gen": {"bleu_acc": 0.0, "bleu_acc_stderr": 0.0, "bleu_diff": 0.0, "bleu_diff_stderr": 0.0, "bleu_max": 0.0, "bleu_max_stderr": 0.0, "bleurt_acc": 0.835985312117503, "bleurt_acc_stderr": 0.012962704327492454, "bleurt_diff": 0.14077322143090107, "bleurt_diff_stderr": 0.005459888909582694, "bleurt_max": -1.4399358725752065, "bleurt_max_stderr": 0.0022126992369197133, "rouge1_acc": 0.0, "rouge1_acc_stderr": 0.0, "rouge1_diff": 0.0, "rouge1_diff_stderr": 0.0, "rouge1_max": 0.0, "rouge1_max_stderr": 0.0, "rouge2_acc": 0.0, "rouge2_acc_stderr": 0.0, "rouge2_diff": 0.0, "rouge2_diff_stderr": 0.0, "rouge2_max": 0.0, "rouge2_max_stderr": 0.0, "rougeL_acc": 0.0, "rougeL_acc_stderr": 0.0, "rougeL_diff": 0.0, "rougeL_diff_stderr": 0.0, "rougeL_max": 0.0, "rougeL_max_stderr": 0.0}}, "versions": {"truthfulqa_gen": 1}}

tests/testdata/truthfulqa_mc-v0-loglikelihood

-226a6783976177dc9ceda5688623ff37023242eff30ddf270b886bf7b9b32228
\ No newline at end of file
+226a6783976177dc9ceda5688623ff37023242eff30ddf270b886bf7b9b32228

tests/testdata/truthfulqa_mc-v0-res.json

-{"results": {"truthfulqa_mc": {"mc1": 0.2141982864137087, "mc1_stderr": 0.01436214815569045, "mc2": 0.465436996173817, "mc2_stderr": 0.0048422530880316405}}, "versions": {"truthfulqa_mc": 0}}
\ No newline at end of file
+{"results": {"truthfulqa_mc": {"mc1": 0.2141982864137087, "mc1_stderr": 0.01436214815569045, "mc2": 0.465436996173817, "mc2_stderr": 0.0048422530880316405}}, "versions": {"truthfulqa_mc": 0}}

tests/testdata/truthfulqa_mc-v1-loglikelihood

-1e07020e9cf41d46ed65312eb39d2b8e6599673d4f0d6b67c0d0eba0efb493bb
\ No newline at end of file
+1e07020e9cf41d46ed65312eb39d2b8e6599673d4f0d6b67c0d0eba0efb493bb

tests/testdata/truthfulqa_mc-v1-res.json

-{"results": {"truthfulqa_mc": {"mc1": 0.23255813953488372, "mc1_stderr": 0.01478915753108052, "mc2": 0.4462325560722362, "mc2_stderr": 0.004986523944692003}}, "versions": {"truthfulqa_mc": 1}}
\ No newline at end of file
+{"results": {"truthfulqa_mc": {"mc1": 0.23255813953488372, "mc1_stderr": 0.01478915753108052, "mc2": 0.4462325560722362, "mc2_stderr": 0.004986523944692003}}, "versions": {"truthfulqa_mc": 1}}