Ecosyste.ms: OpenCollective

An open API service for software projects hosted on Open Collective.

approach0

Approach0: A math-aware search engine.
Collective - Host: opensource - https://opencollective.com/approach0 - Website: https://approach0.xyz - Code: https://github.com/approach0

better indent and show scores in splade_visualize()

github.com/approach0/pya0 - f5738ba3ff1aace7495fbc4e2dea32fecd2a45d3 authored almost 2 years ago by t-k- <[email protected]>
FIX: move unwrap_isolated_tex_groups() into preprocess_for_transformer() and set it and tok_ver=v3 as default.

github.com/approach0/pya0 - aa6425ba1fb9c865fca45ebc75388291bec4d0ea authored almost 2 years ago by t-k- <[email protected]>
show get_dynamic_lr() in colbert.

github.com/approach0/pya0 - 98aadd2400178950c3ab5b6427813d251b536158 authored almost 2 years ago by t-k- <[email protected]>
utils/transformer_eval.ini: added search_arqmath3_task2_cotmae.

github.com/approach0/pya0 - 5f50be2a21420a46b4326cd981bd5cd5576fba58 authored almost 2 years ago by t-k- <[email protected]>
utils/transformer_eval.ini: added index_arqmath3_task2_cotmae

github.com/approach0/pya0 - 77fbaf515c21e82e054f673442af723e06f9fb30 authored almost 2 years ago by t-k- <[email protected]>
update ini file.

github.com/approach0/pya0 - 38df70f284e1ad41c8816ff707e279f30ad1ed10 authored almost 2 years ago by t-k- <[email protected]>
added partial-pretrained experiments for search as well.

github.com/approach0/pya0 - f0fbbc12ac4ea8e21c95b2343d81c0e9b7bbf537 authored almost 2 years ago by t-k- <[email protected]>
update utils/transformer_eval.ini: added partial-pretrained experiments.

github.com/approach0/pya0 - 8394d8b6fe3e4db789434eadb1f23503e4b554d9 authored about 2 years ago by t-k- <[email protected]>
update utils/transformer_eval.ini

github.com/approach0/pya0 - 5a186ecbcc3bf26d84d51ba28ec15f806af7d0c0 authored about 2 years ago by t-k- <[email protected]>
splade class added disable_unused_parameter_for_producing_loss()

github.com/approach0/pya0 - b11a7176f28a5fa6b86a8f21fd494db0f4c28807 authored about 2 years ago by t-k- <[email protected]>
minor change.

github.com/approach0/pya0 - 84d5befeeb66b6a103fa87b61fbcfc5c9e4ced54 authored about 2 years ago by t-k- <[email protected]>
warning log level at model loading stage, and error log level at training.

github.com/approach0/pya0 - 8cc84586409979d2b69be17d1293e561f6ff072f authored about 2 years ago by t-k- <[email protected]>
FIX: SpladeMaxEncoder and DprEncoder from_pretrained() method.

github.com/approach0/pya0 - 5966a20acc2a02b695f55e9f1cf0d0a1d562b7e1 authored about 2 years ago by t-k- <[email protected]>
transformer_utils.py: visualize splade scores.

github.com/approach0/pya0 - 1880ecd7c876803697a371729ae8abced6027c5d authored about 2 years ago by t-k- <[email protected]>
+transformers.logging.set_verbosity_warning()

github.com/approach0/pya0 - 9cb496c3fa3480252bea1fc26599270e875da566 authored about 2 years ago by t-k- <[email protected]>
minor change.

github.com/approach0/pya0 - da017cf3835d21a954630b36fa31b638e394dd3b authored about 2 years ago by t-k- <[email protected]>
added splade_visualize.

github.com/approach0/pya0 - baf7b93a5657aae42c8dc7b6170b063ce74ec383 authored about 2 years ago by t-k- <[email protected]>
a few minor changes.

github.com/approach0/pya0 - 24bb60a526307c0a4eac0ffbffa7330313177de8 authored about 2 years ago by t-k- <[email protected]>
add flops regularizer and fixups.

github.com/approach0/pya0 - f49b97615a8af8f4054425d4af77a236b6ec21b4 authored about 2 years ago by t-k- <[email protected]>
support splade-v2 training.

github.com/approach0/pya0 - e119269269c14029f587a61ed4f557ee40b2008e authored about 2 years ago by t-k- <[email protected]>
transformer_eval.ini: added search_arqmath3_*_dpr for condenser/cotmae

github.com/approach0/pya0 - c2857850e21168239de0262f630c8ff5f80f424a authored about 2 years ago by t-k- <[email protected]>
utils/transformer_eval.ini: added index_arqmath3_condenser_dpr.

github.com/approach0/pya0 - d17816c2f47120a4f564d45b0cfd197e31ed6339 authored about 2 years ago by t-k- <[email protected]>
minor fixup.

github.com/approach0/pya0 - 47b00fe6a909e73c739a71f5838d2cae2cf358ab authored about 2 years ago by t-k- <[email protected]>
update utils/transformer_eval.ini for index_arqmath3_cotmae_dpr.

github.com/approach0/pya0 - cb8c0ccbe0508831b7e5495da484d413da029b77 authored about 2 years ago by t-k- <[email protected]>
minor: print dynamic lr in DPR training.

github.com/approach0/pya0 - 0310ef98fa460e96269cd14711b4babe981b9413 authored about 2 years ago by t-k- <[email protected]>
update arqmath-2021-train-data.py for replace_isolated_groups and num_tokenizer_ver.

github.com/approach0/pya0 - 0971e40117d9ceda552510b538c5aecdbb5c299c authored about 2 years ago by t-k- <[email protected]>
fix pretrain_test print for condenser.

github.com/approach0/pya0 - 92fdaa5422e55933f8f90650e980ceb069676e82 authored about 2 years ago by t-k- <[email protected]>
fix condenser training from ckpt and mae test_loop.

github.com/approach0/pya0 - 9dbde4593478e56aaa8c3ebc81d778f33d1f73f2 authored about 2 years ago by t-k- <[email protected]>
Condenser reproduced decoder outputs -- add .eval() to decoder nn.ModuleList()

github.com/approach0/pya0 - 3ab3f3a2282fee838bfdb1e7a62fec9466c0b166 authored about 2 years ago by t-k- <[email protected]>
fix Condenser loading from pretrained.

github.com/approach0/pya0 - 10a785c0c57222186997682e6a45681e912c2431 authored about 2 years ago by t-k- <[email protected]>
minor change: adjust cross-encoder reranker batch size.

github.com/approach0/pya0 - d8682e6794f34c354abf1b7bd4cd16731f7d6355 authored about 2 years ago by t-k- <[email protected]>
complete requirements.txt

github.com/approach0/pya0 - 25cb4616d5d81c5a751dfaf36bae0aeab9476e91 authored about 2 years ago by t-k- <[email protected]>
added math_10 cross encoder reranking.

github.com/approach0/pya0 - d446c697be045ce77641d6534019dc5a52011798 authored about 2 years ago by t-k- <[email protected]>
Cot-MAE is working.

github.com/approach0/pya0 - 52bb248e1a43b161f0dd65ec359c9439485551e7 authored about 2 years ago by t-k- <[email protected]>
Update README.md

github.com/approach0/math-dense-retrievers - a2c1349884169e2c30861eff4560d38008930325 authored about 2 years ago by Wei <[email protected]>
test_loop adds condenser CTX visualization.

github.com/approach0/pya0 - 126e72716685553a6628e038fb555ed859ba09a6 authored about 2 years ago by t-k- <[email protected]>
utils/transformer.py: test_loop classifier_hook supports condenser MLM for both encoder and decoder.

github.com/approach0/pya0 - ec01d486fab3a1cc28fe167fc6c6317eaed93029 authored about 2 years ago by t-k- <[email protected]>
fix condenser.py, e.g., (p.requires_grad = False) target layer

github.com/approach0/pya0 - 9611ea36b21bc2998da565b1ab904a1bb0435df3 authored about 2 years ago by t-k- <[email protected]>
set Condenser skip_from=6 by default.

github.com/approach0/pya0 - 4ccb6e203ca1b90df544631faba40c7224aff15c authored about 2 years ago by t-k- <[email protected]>
update tests/transformer_unmask.txt for condensor testing.

github.com/approach0/pya0 - 6adf17189d323e919c4cda4336bbef4a0d32d956 authored about 2 years ago by t-k- <[email protected]>
finish dec_ctx_loss calculation, start training again.

github.com/approach0/pya0 - 7308baccfe6fe33d5139fc943cad734aa9382cb8 authored about 2 years ago by t-k- <[email protected]>
minor fixup.

github.com/approach0/pya0 - ad5fc1634f823ddac4a1e1568b5046a142bf5c53 authored about 2 years ago by t-k- <[email protected]>
added copy_weights() and refactor utils/condenser.py

github.com/approach0/pya0 - 8e0f8aae91e28b899b87ca833253eea7b6ab4da6 authored about 2 years ago by t-k- <[email protected]>
fixup, forgot to import re;

github.com/approach0/pya0 - aeeaaaf7ba4d3450933c97d3d258f86a1ebb27da authored about 2 years ago by t-k- <[email protected]>
added replace_isolated_groups option in mse_aops_dataloader()

github.com/approach0/pya0 - ac6ea88ff3cdb4f421d364be5e2541d421de59bb authored about 2 years ago by t-k- <[email protected]>
added unwrap_isolated_tex_groups()

github.com/approach0/a0-crawlers - 5805e09db3ad9acaa2614196d42dd4285ea946f8 authored about 2 years ago by t-k- <[email protected]>
dataset-adapter/mse-aops-2021-train-data.py: data generator for condenser pretrain.

github.com/approach0/pya0 - 0c1802231b9eaf6422f7613aa07a3b991555bd88 authored about 2 years ago by t-k- <[email protected]>
CLS embs remove pooling.

github.com/approach0/pya0 - 894b6045a3380a672887656746a5121cc5e41678 authored about 2 years ago by t-k- <[email protected]>
Update TODO.md

github.com/approach0/docs - 2eeaf7954228e93f57e0486e357a4c490f2464ef authored about 2 years ago by Wei <[email protected]>
fixup: miscalc warmup_steps.

github.com/approach0/pya0 - 87d8366940ac57c2e3cd28652d0bfffda7850cae authored about 2 years ago by t-k- <[email protected]>
change default lr=1e-4

github.com/approach0/pya0 - 29b814c209e890cde45250d49ef49132f19424c3 authored about 2 years ago by t-k- <[email protected]>
added warmup learning rate.

github.com/approach0/pya0 - 891f4bcbf412af21a5cacc6c77d5a45cfcbc4c1d authored about 2 years ago by t-k- <[email protected]>
plug in condenser for pretraining.

github.com/approach0/pya0 - 1130f85cd003142774f8ff2dd7db6090f91c156b authored about 2 years ago by t-k- <[email protected]>
indent fixup

github.com/approach0/math-dense-retrievers - bf8e364e8ab6cbb2ead6b2f1f049371f0f7eea6a authored about 2 years ago by t-k- <[email protected]>
update README for condenser training.

github.com/approach0/math-dense-retrievers - ac9efcb58a351ec85511fb8d2159bec75d7e1f5a authored about 2 years ago by t-k- <[email protected]>
added Condenser/MAE loss compute.

github.com/approach0/pya0 - 734d3b13f0a8d615c24d2bd3ec118c08f68b4e1b authored about 2 years ago by t-k- <[email protected]>
added utils/condenser.py

github.com/approach0/pya0 - b72a3442a45a2f7cf19f6040fc1e0375842b978a authored about 2 years ago by t-k- <[email protected]>
preprocess.sh: remove colons in runfile names.

github.com/approach0/pya0 - 6a35538567c298a0e2ebca0e8d72b54dd54c8d18 authored about 2 years ago by t-k- <[email protected]>
minor change.

github.com/approach0/pya0 - ddb8e03487442dfe711ddde0c6fc8dd3c12da6b5 authored about 2 years ago by t-k- <[email protected]>
update eval-ntcir12.sh

github.com/approach0/pya0 - 68e49c2c41e5ebe025b2c1e05505f3849308c67a authored about 2 years ago by t-k- <[email protected]>
added effectiveness_ttest.py

github.com/approach0/pya0 - b351f9be966d80c28ad532e02ec339335b648c16 authored about 2 years ago by t-k- <[email protected]>
Create README.md

github.com/approach0/open_research_objectives - 45e871fb8cf208e61efd66a22ab6b6954e7b37c7 authored about 2 years ago by Wei <[email protected]>
README: update training command.

github.com/approach0/math-dense-retrievers - de996904cf4065f154e279f82b8375673cbc731b authored about 2 years ago by t-k- <[email protected]>
refine pft_print and tests/transformer_unmask.txt

github.com/approach0/pya0 - 1674174e9d2810938d7a7f979501ea25576ee404 authored about 2 years ago by t-k- <[email protected]>
README: add instruction for test.txt create.

github.com/approach0/math-dense-retrievers - dfa7a965bd138bce78b25ee8a50f58ff2605c270 authored about 2 years ago by t-k- <[email protected]>
update README for more training data creation instructions.

github.com/approach0/math-dense-retrievers - cbdf0cbaeae916d0efb374ac86802234070e9354 authored about 2 years ago by t-k- <[email protected]>
minor adjustment in print freq.

github.com/approach0/pya0 - 57120f089f359a0df5bdf65e1401ee6657706482 authored about 2 years ago by t-k- <[email protected]>
fixup again.

github.com/approach0/pya0 - 22ff16943c1ecc428c203d853fbe7b8c2b069bf1 authored about 2 years ago by t-k- <[email protected]>
fixup num_tokenizer_ver=3

github.com/approach0/pya0 - e6ee9e8abd5933a779744a01c51aa4eb961ce75d authored about 2 years ago by t-k- <[email protected]>
dataset-adapter/mse-aops-2021.py: switch to num_tokenizer v3.

github.com/approach0/pya0 - 9ea3fe1854c2d1d86cbaddb2f6b70f5b6303c864 authored about 2 years ago by t-k- <[email protected]>
found bug in num_tokenizer_v2, added num_tokenizer_v3.

github.com/approach0/pya0 - 0641edb4adeb95ad303fb364de794521bdf37527 authored about 2 years ago by t-k- <[email protected]>
finish implementing colbert visualizer.

github.com/approach0/pya0 - 6348cd936775d862fd781c38c13aa9042eab97b8 authored about 2 years ago by t-k- <[email protected]>
preparing hit vis_* info for html_output.

github.com/approach0/pya0 - 3de8095d51e0ac1937a6002d53634bde25e9dc00 authored about 2 years ago by t-k- <[email protected]>
writing colbert run visualizer.

github.com/approach0/pya0 - a330ab3bf309f8287c8bff38ac642660415f4d11 authored about 2 years ago by t-k- <[email protected]>
minor change for colbert_visualize()

github.com/approach0/pya0 - 186def2daf12f1388bd94aa79a8e68ec70def5d6 authored about 2 years ago by t-k- <[email protected]>
minor: remove redundant line.

github.com/approach0/pya0 - f0c22375d53e13ab25eee680f8c2e2e0e7c08435 authored about 2 years ago by t-k- <[email protected]>
filter out wildcards in qrels.ntcir12-math-browsing-concrete.txt

github.com/approach0/pya0 - 138f7543e814ab202169ebb4aa71e90e7694a770 authored over 2 years ago by t-k- <[email protected]>
rename experiments/arqmath3-a0setup.sh -> experiments/arqmath3-setup.sh

github.com/approach0/pya0 - e0e6a80720fe92d487c927d1f6254319d262b1be authored over 2 years ago by t-k- <[email protected]>
added experiments/arqmath3-a0setup.sh

github.com/approach0/pya0 - c64f296bd9e325ce051d772a6c76aa27a4ac2de6 authored over 2 years ago by t-k- <[email protected]>
Update vdisk.md

github.com/approach0/docs - b80b2f154065a3a971028cab4f750a58ed3da135 authored over 2 years ago by Wei <[email protected]>
Update appendix.md

github.com/approach0/docs - 87bbf8bf173ed9ad9e396f14f5b6115020b92869 authored over 2 years ago by Wei <[email protected]>
Update appendix.md

github.com/approach0/docs - 7b83def2e75a8a6c9eb1198c6cc1cda16fde46cb authored over 2 years ago by Wei <[email protected]>
added eval-arqmath3/task3

github.com/approach0/pya0 - f2852c9c508bfc8357fd40d929cbbf2749c89fca authored over 2 years ago by t-k- <[email protected]>
added eval-arqmath3/task2 (and qrel file)

github.com/approach0/pya0 - 6f2b99414776c4825c6f1b6d5a3e7334da9d79cf authored over 2 years ago by t-k- <[email protected]>
replace and align eval-arqmath2-task2 (also new qrel file)

github.com/approach0/pya0 - 7916475dc17cc5385cce563c6504bc27038de239 authored over 2 years ago by t-k- <[email protected]>
added bpref and judge_rate metrics for eval-arqmath1-task2

github.com/approach0/pya0 - 1da030a6dc3b7a0161ecaf7d53871a852c404d5c authored over 2 years ago by t-k- <[email protected]>
added cache for de_duplicate.py and QREL guess for eval.sh.

github.com/approach0/pya0 - cbc6d91b3c25b00ed4cd973152d8207781b60c58 authored over 2 years ago by t-k- <[email protected]>
aligned official results with eval-arqmath1-task2.

github.com/approach0/pya0 - 5867a201b853b2c863175fa7da8d2f83a8364275 authored over 2 years ago by t-k- <[email protected]>
added qrels.arqmath-2020-task2-visual_ids.v3.txt

github.com/approach0/pya0 - cbd82113de48e40d369a20ab1e60dd512ede811c authored over 2 years ago by t-k- <[email protected]>
update eval-arqmath1-task2.

github.com/approach0/pya0 - 303462298fea7e743875424c19e627ea9bbe2fe8 authored over 2 years ago by t-k- <[email protected]>
added eval-arqmath3/task1

github.com/approach0/pya0 - 191890f8287363a646570ed96ef06331977e58de authored over 2 years ago by t-k- <[email protected]>
added topics-and-qrels/qrels.arqmath-2022-task1-official.txt

github.com/approach0/pya0 - 736833c369a8b36f467a0c75b3a623fe3438eeba authored over 2 years ago by t-k- <[email protected]>
fix eval-arqmath2-task1 to support space in file names.

github.com/approach0/pya0 - 5cabca62bff790422ab58379690e8842a9b82760 authored over 2 years ago by t-k- <[email protected]>
fix eval-arqmath1-task1.sh to support space in file names.

github.com/approach0/pya0 - c15949c4b4e7384d2fc198945eac68ed056441d4 authored over 2 years ago by t-k- <[email protected]>
Update experiments.tar.gz Vault HASH

github.com/approach0/math-dense-retrievers - 3517331a397823fb185682468f17a07e327f3b1b authored over 2 years ago by Wei <[email protected]>
Update README.md

github.com/approach0/math-dense-retrievers - d79a035e716d99fdc245673478fad3bcaf9e52d0 authored over 2 years ago by Wei <[email protected]>
minor file move

github.com/approach0/pya0 - 29dd30ba988aa004353cca7fee561bd828cb814d authored over 2 years ago by t-k- <[email protected]>
fix tests/genn-plot.py, modify pya0.parse().

github.com/approach0/pya0 - c12f8a478b32f57c5ddc0d09f8c9168a1065eb3f authored over 2 years ago by t-k- <[email protected]>
update utils/transformer_eval.ini

github.com/approach0/pya0 - 24c2f435c74b2c4a2bb13fed88a239669f461e69 authored over 2 years ago by t-k- <[email protected]>
generate detailed ColBERT time report.

github.com/approach0/pya0 - e0c1610f66c824bdc6afb3ba6f773642066c1a6c authored over 2 years ago by t-k- <[email protected]>