Commit Graph

112 Commits

Author SHA1 Message Date
Hstellar c5fe609bd4 Adding initial hyper-parameter tuning file 2022-06-08 02:30:28 -04:00
Kashif Rasul 019a0a391c Merge pull request #2 from kashif/switch
initial switch transformer
2022-06-06 14:37:16 +02:00
Kashif Rasul 6cbce657e1 notebook 2022-06-06 14:36:10 +02:00
Kashif Rasul d11578a235 fix final_output tensor shape 2022-06-06 14:11:51 +02:00
Kashif Rasul 0b0ecc94f9 fix typo 2022-06-06 11:45:31 +02:00
Kashif Rasul c99572a6d9 fix args 2022-06-06 11:44:19 +02:00
Kashif Rasul 159254348b fix notebook 2022-06-06 11:42:52 +02:00
Kashif Rasul 3eeabcd1d8 import 2022-06-06 11:34:14 +02:00
Kashif Rasul a429cef13d added switch encoder layer 2022-06-06 11:04:43 +02:00
Kashif Rasul 40981abdb4 initial switch transformer 2022-06-06 10:35:36 +02:00
Kashif Rasul 39e613be84 use torch.no_grad 2022-05-31 11:28:46 +02:00
Kashif Rasul 0e8d816aac fixed xformers now getting runtime error 2022-05-30 11:50:33 +02:00
Kashif Rasul 8b62f7194c fix torch.topk use 2022-05-28 18:49:50 +02:00
Kashif Rasul 8f52119ea6 fix device runs on cpu now 2022-05-27 08:49:13 +02:00
Kashif Rasul 0dadc25c3a added encode function 2022-05-25 11:04:12 +02:00
Kashif Rasul b65415bbe2 added encoder-xformer-vanilla transformer 2022-05-24 12:57:17 +02:00
Kashif Rasul 335f96acd7 sym scaling 2022-05-21 21:02:21 +02:00
Kashif Rasul bc5c400713 fix typos 2022-05-21 11:42:35 +02:00
Kashif Rasul cfc8daa4fa added decoder scaling plot 2022-05-21 11:13:00 +02:00
Kashif Rasul af6b4ea765 added dec and sym exp loops 2022-05-20 14:50:14 +02:00
Kashif Rasul 8628813bae remove output 2022-05-20 14:31:03 +02:00
Kashif Rasul cf166a1b53 some plots 2022-05-20 14:29:59 +02:00
Kashif Rasul c1f930658a initial experiment 2022-05-20 11:05:30 +02:00
Kashif Rasul adb975c199 added s4 requirements 2022-05-10 10:58:03 +02:00
Kashif Rasul b2e37ef867 added s4 2022-05-10 10:52:14 +02:00
Kashif Rasul 8e36564b62 initial inference 2022-04-29 16:08:21 +02:00
Kashif Rasul 1bbb48c6a3 fixed names 2022-04-29 13:28:08 +02:00
Kashif Rasul 20217bc9c8 initial reformer 2022-04-29 12:54:58 +02:00
Kashif Rasul a88428b254 remove .DS_Store 2022-04-29 09:13:06 +02:00
Hstellar 8da5cb49d9 cleaned up(still not working) 2022-04-20 00:57:39 -04:00
Hstellar c7afec7ac5 Clean up(still not working) 2022-04-19 16:29:30 -04:00
Kashif Rasul d72782f689 train for 50 epochs 2022-04-13 21:06:45 +02:00
Kashif Rasul c280009c0c fix case when L_Q or L_K is 1 2022-04-13 20:53:03 +02:00
Kashif Rasul 3f5bae3c73 added .typesafe 2022-04-13 14:51:46 +02:00
Kashif Rasul a8a046ab0d added informer 2022-04-13 14:51:21 +02:00
Kashif Rasul 07834941c2 remove 2022-04-13 11:42:45 +02:00
Kashif Rasul 1b07a5050e rename 2022-04-13 11:39:40 +02:00
Kashif Rasul f02d4f801b remove repeated arg 2022-04-13 09:49:23 +02:00
Kashif Rasul 82eb1ca3a5 formatting 2022-04-13 09:46:29 +02:00
Hstellar 3e3f46bec4 Adding pyraformer(not working) 2022-04-12 13:28:37 -04:00
Kashif Rasul 38882af8aa added ETSFormer 2022-04-07 13:29:56 +02:00
Kashif Rasul bc331ba3a7 Update README.md 2022-04-06 13:45:56 +02:00
Kashif Rasul 944a697289 ignore lightning_logs 2022-04-06 13:28:59 +02:00
Kashif Rasul b3bd636a7c move template 2022-04-06 13:27:02 +02:00
Kashif Rasul ed4091620d added notebook 2022-04-06 13:24:40 +02:00
Kashif Rasul 6f64227d31 added autoformer predict 2022-04-06 13:15:58 +02:00
Kashif Rasul 68d33b7fd3 add estimator args 2022-04-04 22:21:04 +02:00
Kashif Rasul bd0165f720 initial enc dec output 2022-04-04 22:06:17 +02:00
Kashif Rasul c1df7b97f1 initial autoformer
not working
2022-04-04 16:54:13 +02:00
Kashif Rasul 51b7854181 fix default activation 2022-04-04 15:25:01 +02:00