{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":499127120,"defaultBranch":"master","name":"few-shot-hypernets-public","ownerLogin":"gmum","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2022-06-02T12:30:33.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/5064513?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1712059370.0","currentOid":""},"activityList":{"items":[{"before":"613a186bd165402a08195889e016d000ee728f3f","after":"62a2f860e048187f255e562040fcbf8c4b81a735","ref":"refs/heads/master","pushedAt":"2024-07-11T21:37:14.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"BayesHyperShot (#7)\n\n* bruh\r\n\r\n* hn output print\r\n\r\n* best model plots, bayeslinear1 support\r\n\r\n* eksperyment xd\r\n\r\n* FIX\r\n\r\n* log sigma and mu in experiment\r\n\r\n* fix\r\n\r\n* Handle histograms etc\r\n\r\n* Fix\r\n\r\n* remove redundant reparam\r\n\r\n* Improve plots\r\n\r\n* Better plots\r\n\r\n* Revert \"Better plots\"\r\n\r\nThis reverts commit bae46d0f62e9f31f55a6265b55e474a9d3002e77.\r\n\r\n* Add separator\r\n\r\n* no reparam scalling\r\n\r\n* sigma norm loss test\r\n\r\n* quickfix\r\n\r\n* qf #2\r\n\r\n* .\r\n\r\n* Add constant dkl scaler\r\n\r\n* Cleanup PR\r\n\r\n* Change flags naming conventions\r\n\r\n* typo fix\r\n\r\n* Typo fix\r\n\r\n* Typo fix\r\n\r\n* Add epoch state dict for reparam scaling\r\n\r\n* Remove param that does not exist\r\n\r\n* fix\r\n\r\n* fix\r\n\r\n* fix\r\n\r\n* fix\r\n\r\n* Add reparam flags\r\n\r\n* Use translation\r\n\r\n* fix\r\n\r\n* fix\r\n\r\n* fix\r\n\r\n* bruh\r\n\r\n* :)\r\n\r\n* Add bayesian target network on testing\r\n\r\n* Delete reparam scaling from BayesLinear.forward\r\n\r\n* Add skeleton logic for saving and updating testing mu and sigma on testing\r\n\r\n* Add missing logic for logging bayesian target network params on test\r\n\r\n* Add sampling on test\r\n\r\n* -4\r\n\r\n* add temporary print\r\n\r\n* add temporary print\r\n\r\n* add printing helper in meta template\r\n\r\n* add printing helper in meta template\r\n\r\n* print helper\r\n\r\n* print helper\r\n\r\n* Add sampling in another set_forward\r\n\r\n* Modify another set_forward_with_adaptation\r\n\r\n* fix bug with query accuracy scores\r\n\r\n* print helper\r\n\r\n* improve testing\r\n\r\n* Helper printer\r\n\r\n* Simplify uploading to neptune\r\n\r\n* Fix printing\r\n\r\n* Fix bug with histograms on train\r\n\r\n* print every 100 epochs\r\n\r\n* exp no train\r\n\r\n* Add HyperShot experiment\r\n\r\n* Add args loading\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Setup new method of generating params for experiment\r\n\r\n* Fix of bash script\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Improve param generation\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Change formatting of params\r\n\r\n* Add parser creator\r\n\r\n* Skip default params\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Change formatting of checkpoint suffix\r\n\r\n* Fix typo\r\n\r\n* Add instruction\r\n\r\n* Fix typo\r\n\r\n* Add experiment template\r\n\r\n* Format instruction\r\n\r\n* Format template and instruction\r\n\r\n* Experiment done\r\n\r\n* fix def\r\n\r\n* qfix\r\n\r\n* qfix2\r\n\r\n* qfix3\r\n\r\n* qfix4\r\n\r\n* qfix5\r\n\r\n* gfix6\r\n\r\n* types...\r\n\r\n* bruh\r\n\r\n* pls just work\r\n\r\n* :c\r\n\r\n* :')\r\n\r\n* hello\r\n\r\n* dimension fix?\r\n\r\n* .\r\n\r\n* plox\r\n\r\n* .\r\n\r\n* .\r\n\r\n* pls work wuw\r\n\r\n* :C\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* ?\r\n\r\n* types\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* fix?\r\n\r\n* debug\r\n\r\n* types...\r\n\r\n* :)\r\n\r\n* .\r\n\r\n* stupid mistake\r\n\r\n* .\r\n\r\n* pls work\r\n\r\n* types\r\n\r\n* .\r\n\r\n* cuda\r\n\r\n* .\r\n\r\n* dim\r\n\r\n* :D\r\n\r\n* debug\r\n\r\n* debug2\r\n\r\n* debug3\r\n\r\n* concat\r\n\r\n* fix\r\n\r\n* misspell\r\n\r\n* .\r\n\r\n* fxd?\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* build relations features\r\n\r\n* q reshape\r\n\r\n* neptune debug\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* test full dataset\r\n\r\n* .\r\n\r\n* .\r\n\r\n* fallback\r\n\r\n* pls worky\r\n\r\n* ?\r\n\r\n* fix\r\n\r\n* 100\r\n\r\n* ?\r\n\r\n* 30\r\n\r\n* debug\r\n\r\n* shape\r\n\r\n* Print loading stages\r\n\r\n* Add printing of ARGSPATH\r\n\r\n* Print parsed args\r\n\r\n* Delete redundant print\r\n\r\n* Get rid of neptune infra\r\n\r\n* Print S and Q shapes\r\n\r\n* test\r\n\r\n* Fix\r\n\r\n* Debug\r\n\r\n* Reshape y\r\n\r\n* zip dataset with targets\r\n\r\n* split targets\r\n\r\n* Continiue splitting\r\n\r\n* Fix typo\r\n\r\n* Move computation to GPU\r\n\r\n* Get to separate tasks\r\n\r\n* cast zip to list\r\n\r\n* Fix typo\r\n\r\n* Log tensor values\r\n\r\n* Add proper forward of query set\r\n\r\n* Fix\r\n\r\n* Delete batches\r\n\r\n* Fix\r\n\r\n* Print shape\r\n\r\n* Fix\r\n\r\n* Fix\r\n\r\n* Delete redundant dim\r\n\r\n* Add comments\r\n\r\n* Change comments\r\n\r\n* Change comments\r\n\r\n* done\r\n\r\n* copy\r\n\r\n* idxs\r\n\r\n* dims\r\n\r\n* debug\r\n\r\n* tensors...\r\n\r\n* dim=-1\r\n\r\n* flatten [1x5]\r\n\r\n* fix hist\r\n\r\n* colors, 100\r\n\r\n* comma\r\n\r\n* 200\r\n\r\n* 200\r\n\r\n* 1000\r\n\r\n* 250\r\n\r\n* final histograms\r\n\r\n* visual changes\r\n\r\n* .\r\n\r\n* align\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* seaborn\r\n\r\n* melt\r\n\r\n* .\r\n\r\n* typO\r\n\r\n* brah\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* .\r\n\r\n* Clean code\r\n\r\n* Merge\r\n\r\n* reconciling hs and maml\r\n\r\n* fixes for hs when neptune not avialable\r\n\r\n* add bayesian hypershot commands\r\n\r\n* update readme\r\n\r\n---------\r\n\r\nCo-authored-by: OneAndZero24 <oneandzero24@gmail.com>\r\nCo-authored-by: Mateusz Rajski <60450261+mateuuszzzzz@users.noreply.github.com>\r\nCo-authored-by: mateuuszzzzz <mati.icloud2000@gmail.com>\r\nCo-authored-by: Mateusz Rajski <mateuszrajski@Mateuszs-MacBook-Air.local>\r\nCo-authored-by: mateuuszzzzz <mateusz.rajski@student.uj.edu.pl>","shortMessageHtmlLink":"BayesHyperShot (<a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2404167054\" data-permission-text=\"Title is private\" data-url=\"https://github.com/gmum/few-shot-hypernets-public/issues/7\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/gmum/few-shot-hypernets-public/pull/7/hovercard\" href=\"https://github.com/gmum/few-shot-hypernets-public/pull/7\">#7</a>)"}},{"before":"d75ad19409ed477c9c1028930a0441f7211a019a","after":"613a186bd165402a08195889e016d000ee728f3f","ref":"refs/heads/master","pushedAt":"2024-04-29T11:10:03.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"Add files via upload","shortMessageHtmlLink":"Add files via upload"}},{"before":"a439e616d74b5e2b46edeadffc087ff1960faf52","after":"97be75efe96a857a7c87430aab2873403c48246c","ref":"refs/heads/state-before-forking-for-bayesianhypershot","pushedAt":"2024-04-02T13:03:34.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"update readme","shortMessageHtmlLink":"update readme"}},{"before":"a439e616d74b5e2b46edeadffc087ff1960faf52","after":null,"ref":"refs/heads/merge-bayesian-hypershot","pushedAt":"2024-04-02T12:02:50.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"}},{"before":null,"after":"a439e616d74b5e2b46edeadffc087ff1960faf52","ref":"refs/heads/merge-bayesian-hypershot","pushedAt":"2024-04-02T12:02:05.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"add bayesian hypershot commands","shortMessageHtmlLink":"add bayesian hypershot commands"}},{"before":"83bd1870111f3cee2e81cbf10b727f21471cef1b","after":"a439e616d74b5e2b46edeadffc087ff1960faf52","ref":"refs/heads/state-before-forking-for-bayesianhypershot","pushedAt":"2024-04-02T12:01:33.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"add bayesian hypershot commands","shortMessageHtmlLink":"add bayesian hypershot commands"}},{"before":"d3f4a966ea5b82f56eaffb7121c09ec6be3d0dc4","after":"83bd1870111f3cee2e81cbf10b727f21471cef1b","ref":"refs/heads/state-before-forking-for-bayesianhypershot","pushedAt":"2024-04-02T11:36:25.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"fixes for hs when neptune not avialable","shortMessageHtmlLink":"fixes for hs when neptune not avialable"}},{"before":"c542fd672e92da08734064f1593b2041291eebbb","after":"d3f4a966ea5b82f56eaffb7121c09ec6be3d0dc4","ref":"refs/heads/state-before-forking-for-bayesianhypershot","pushedAt":"2024-04-02T11:01:34.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"reconciling hs and maml","shortMessageHtmlLink":"reconciling hs and maml"}},{"before":"462600474ceaedc53a6886b5cc98d619d1241a72","after":"c542fd672e92da08734064f1593b2041291eebbb","ref":"refs/heads/state-before-forking-for-bayesianhypershot","pushedAt":"2024-04-02T09:30:44.000Z","pushType":"push","commitsCount":43,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"Merge","shortMessageHtmlLink":"Merge"}},{"before":"e85cc299e0c56a1fe8b2e28ae51e800dfa5595dd","after":"462600474ceaedc53a6886b5cc98d619d1241a72","ref":"refs/heads/state-before-forking-for-bayesianhypershot","pushedAt":"2024-03-23T20:19:53.000Z","pushType":"pr_merge","commitsCount":318,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"Merge pull request #5 from OneAndZero24/master\n\nBayesian HyperShot","shortMessageHtmlLink":"Merge pull request <a class=\"issue-link js-issue-link\" data-error-text=\"Failed to load title\" data-id=\"2204030081\" data-permission-text=\"Title is private\" data-url=\"https://github.com/gmum/few-shot-hypernets-public/issues/5\" data-hovercard-type=\"pull_request\" data-hovercard-url=\"/gmum/few-shot-hypernets-public/pull/5/hovercard\" href=\"https://github.com/gmum/few-shot-hypernets-public/pull/5\">#5</a> from OneAndZero24/master"}},{"before":null,"after":"e85cc299e0c56a1fe8b2e28ae51e800dfa5595dd","ref":"refs/heads/state-before-forking-for-bayesianhypershot","pushedAt":"2024-03-23T20:15:19.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"mprzewie","name":"Marcin Przewięźlikowski","path":"/mprzewie","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/21194406?s=80&v=4"},"commit":{"message":"Merge HyperShot and HyperMAML\n\n* hypernet POC, not yet working\r\n\r\n* 70% acc after fixes\r\n\r\n* to taskset or not to taskset?\r\n\r\n* rm some comments\r\n\r\n* print training acc\r\n\r\n* achieved 80% accuracy after some tweaks\r\n\r\n* 90%!\r\n\r\n* add todos\r\n\r\n* add permutations\r\n\r\n* 70% on crosschar\r\n\r\n* some crazy HN architectures, but no breakthroughs\r\n\r\n* add kernel implementation\r\n\r\n* better kernel(?) architecture, achieved 80.03% on cross-char\r\n\r\n* add metrics dumping\r\n\r\n* make hn_poc configurable via script args\r\n\r\n* update .gitignore to ignore data files\r\n\r\n* gridsearch preparation\r\n\r\n* add an initial version of metrics explorer\r\n\r\n* updates to grid.sh\r\n\r\n* update grid.sh\r\n\r\n* load metrics if resuming\r\n\r\n* measure how hyperparams influence the metrics\r\n\r\n* add neptune and ability to set lr\r\n\r\n* fix resuming\r\n\r\n* resume nept runs\r\n\r\n* new gridsearch\r\n\r\n* new gridsearch\r\n\r\n* fix\r\n\r\n* some new args for detaching\r\n\r\n* some newer args for detaching\r\n\r\n* new gridsearch\r\n\r\n* new gridsearch\r\n\r\n* new gridsearch\r\n\r\n* fixes\r\n\r\n* fixes\r\n\r\n* add layernorm\r\n\r\n* add dropout\r\n\r\n* Attention-based input added\r\n\r\n* simple early stopping\r\n\r\n* lr scheduling\r\n\r\n* lr scheduling\r\n\r\n* small changes - in attention-based approach\r\n\r\n* add import\r\n\r\n* different activations\r\n\r\n* kernel approach to test and set params\r\n\r\n* try out hypnettorch implementation\r\n\r\n* After fitting dimensions\r\n\r\n* Delete trash prints\r\n\r\n* update in the process of building of kernel values matrix\r\n\r\n* more hypnettorch implementations\r\n\r\n* more hypnettorch implementations\r\n\r\n* fix to shmlp\r\n\r\n* fix to shmlp\r\n\r\n* fix to shmlp\r\n\r\n* add chmlp\r\n\r\n* fix\r\n\r\n* basic shmlp\r\n\r\n* non-basic shmlp\r\n\r\n* fix\r\n\r\n* fix shmlp\r\n\r\n* set forward with adaptation\r\n\r\n* set forward with adaptation\r\n\r\n* set forward with adaptation\r\n\r\n* eval freq\r\n\r\n* Add grid scripts\r\n\r\n* Support vs. Support Kernel based method\r\n\r\n* val opt can be sgd\r\n\r\n* script and small changes\r\n\r\n* various pooling strategies, attention after matrix K (supports vs. supports), scalar product instead of NNKernel, using support embeddings after matrix K (supports vs. supports) - ALL: FIRST IMPLEMENTATIONS - TO BE CAREFULLY CHECKED!\r\n\r\n* small fixes after checking the added implementations\r\n\r\n* adapt tn\r\n\r\n* adapt feature+tn\r\n\r\n* small fixes and grid search script\r\n\r\n* fix\r\n\r\n* fix\r\n\r\n* fix args for baselines\r\n\r\n* grid scripts\r\n\r\n* small changes\r\n\r\n* small changes 2\r\n\r\n* small fixes and grid search script\r\n\r\n* adding zeros on the diagonal\r\n\r\n* add cuda\r\n\r\n* add cuda 2\r\n\r\n* remove printing\r\n\r\n* add grids\r\n\r\n* remove zeros from matrix K\r\n\r\n* fix\r\n\r\n* fix 2\r\n\r\n* fix 3\r\n\r\n* ensemble evaluation\r\n\r\n* fix acc names\r\n\r\n* some more lr schedulers\r\n\r\n* change scheduler\r\n\r\n* finetune sup sup kernel\r\n\r\n* fix finetuning\r\n\r\n* grid\r\n\r\n* changes which marcin s made but were untracked\r\n\r\n* finetune sup sup kernel\r\n\r\n* sum aggregation\r\n\r\n* mean agg\r\n\r\n* log cmd in neptune\r\n\r\n* grid cub\r\n\r\n* small fix, self.n_query update in each batch\r\n\r\n* grid cub\r\n\r\n* rm hardcoded n query\r\n\r\n* test.py fixes\r\n\r\n* first implementation of kernel hypothesis test - run it with --use_scalar_product!!!!!!!!!!!!!!!!!\r\n\r\n* add suffix to checkpoint\r\n\r\n* fix\r\n\r\n* full test after train\r\n\r\n* full test after train\r\n\r\n* fix\r\n\r\n* dkt with our best kernel\r\n\r\n* dkt confings change to use nn kernel\r\n\r\n* adding cosine distance instead of scalar product and kernel\r\n\r\n* new kernels\r\n\r\n* fix bkb\r\n\r\n* cosine distance small fix\r\n\r\n* ch1\r\n\r\n* Konrad's changes\r\n\r\n* cross 5 pooling strategy\r\n\r\n* rm permutations in test loop\r\n\r\n* some cleanups\r\n\r\n* dont use query when building embedding\r\n\r\n* dont use query when building embedding\r\n\r\n* fix\r\n\r\n* hack query accuracy\r\n\r\n* hack query accuracy\r\n\r\n* hack query accuracy\r\n\r\n* hack query accuracy\r\n\r\n* hack query accuracy\r\n\r\n* hack query accuracy\r\n\r\n* readd cosine distance\r\n\r\n* separate hypernet_poc into multiple files\r\n\r\n* fix\r\n\r\n* fix\r\n\r\n* fixes\r\n\r\n* perm test\r\n\r\n* reuse query_accuracy\r\n\r\n* unify concatting and few shot strategy\r\n\r\n* unify strategy and concatting\r\n\r\n* fix\r\n\r\n* afs\r\n\r\n* permutation check in hypernet poc\r\n\r\n* rm conv out size\r\n\r\n* rm conv out size\r\n\r\n* analyze which params were used and which not\r\n\r\n* rm fstr\r\n\r\n* new grid on crosschar\r\n\r\n* fix to crosschar\r\n\r\n* cub 1shot\r\n\r\n* refactor code further\r\n\r\n* fix\r\n\r\n* fix params\r\n\r\n* add accuracy of the kernel\r\n\r\n* fix\r\n\r\n* add relational accuracy\r\n\r\n* adding kernel convolutions\r\n\r\n* add PPA\r\n\r\n* readd kernel final dim\r\n\r\n* readd kernel final dim\r\n\r\n* add sup sup PPA\r\n\r\n* grid changes\r\n\r\n* small fixes\r\n\r\n* small fixes 2\r\n\r\n* adding CosineNNKernel\r\n\r\n* grids\r\n\r\n* refactor kernel to implement nn+cosine\r\n\r\n* rm -1 from cosine kernel\r\n\r\n* readd *= -1\r\n\r\n* remove *= -1\r\n\r\n* update cub gridsearch\r\n\r\n* grid scripts\r\n\r\n* readd neg 1\r\n\r\n* test with different hn val epochs\r\n\r\n* save features in train\r\n\r\n* changes for baselines\r\n\r\n* fix\r\n\r\n* fix for beaselines\r\n\r\n* add HyperMAML model for using hypernetworks inside of MAML algorithm\r\n\r\n* hypermaml enhancement of embeddings, delta params regularization\r\n\r\n* add delta parameters saving\r\n\r\n* handling of class embeddings as an input of hypernet\r\n\r\n* alpha parameter for tuning hypernetwork output\r\n\r\n* add milestones parameter\r\n\r\n* calculate loss both on support and query sets\r\n\r\n* add MAML warmup\r\n\r\n* add new features to maml warmup\r\n\r\n* add support for MAML in the backbone models\r\n\r\n* add support for saving MAML feature net\r\n\r\n* enable classifier only MAML training\r\n\r\n* load feature net from file, detach embedding before entering the hypernet\r\n\r\n* add set forward with adaptation\r\n\r\n* skip train and evaluate model\r\n\r\n* hypernet support for multilayer classifier\r\n\r\n* support for generating delta parameters when no per class weight generation is used\r\n\r\n* updates\r\n\r\n* add missing flag when performing full_test, do not save fts for hyper_maml model\r\n\r\n* change RandomSizedCrop to RandomResizedCrop\r\n\r\n* Make class_mean work with class_batch_inputs\r\n\r\n* move back fileleists\r\n\r\n* measuring time\r\n\r\n* update readme\r\n\r\n* remove deprecated models\r\n\r\n* clean up args\r\n\r\nCo-authored-by: Marcin Sendera <marcin.sendera@gmail.com>\r\nCo-authored-by: Marcin Sendera <msendera@github.com>\r\nCo-authored-by: Marcin Przewięźlikowski <marcin.przewiezlikowski@fornax.ai>\r\nCo-authored-by: przem7 <przemyslaw.przybysz97@gmail.com>","shortMessageHtmlLink":"Merge HyperShot and HyperMAML"}}],"hasNextPage":false,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEfUcKsAA","startCursor":null,"endCursor":null}},"title":"Activity · gmum/few-shot-hypernets-public"}