Commit Graph

2144 Commits (3fdf2a63aaf901f16763fa632269b823915199f4)
 

Author SHA1 Message Date
mrq 3fdf2a63aa fixes 2023-03-11 01:18:25 +07:00
mrq b5c6acec9e gutted bloat loggers, now all my useful metrics update per step 2023-03-10 22:34:37 +07:00
mrq bf94744514 I am going to scream 2023-03-09 22:47:46 +07:00
mrq 84c8196da5 Shamelessly nabbed from ae80992817 (if this is makes a big enough difference in training i'm going to cum) 2023-03-09 03:39:23 +07:00
mrq 0ee0f46596 . 2023-03-09 00:29:25 +07:00
mrq 6eb7ebf847 silence printing the model because it's just useless noise 2023-03-04 16:38:24 +07:00
mrq 71cc43e65c added a flag (thanks gannybal) 2023-02-26 14:56:26 +07:00
mrq 0f04206aa2 added ability to toggle some settings with envvars for later testing without needing to manually edit this file (and some other things like disabling it when a user requests it in the future) 2023-02-24 23:08:56 +07:00
mrq 1433b7c0ea working Embedding override 2023-02-23 07:28:27 +07:00
mrq 94aefa3e4c silence 2023-02-23 07:25:09 +07:00
mrq fd66c4104b ugh 2023-02-23 07:18:07 +07:00
mrq 7bcedca771 I guess I can't easily toggle it outside of here, but it works 2023-02-23 07:02:06 +07:00
mrq 0ef8ab6872 shut up 2023-02-23 06:12:27 +07:00
mrq 58600274ac Disabling bitsandbytes optimization as default for now, in the off chance that it actually produces garbage (which shouldn't happen, there's no chance, if training at float16 from a model at float16 works fine, then this has to work) 2023-02-23 03:22:59 +07:00
mrq 918473807f Merge pull request 'bitsandbytes' (#2) from bitsandbytes into master
Reviewed-on: #2
2023-02-23 03:16:25 +07:00
mrq 6676c89c0e I sucked off the hyptothetical wizard again, just using BNB's ADAM optimizer nets HUGE savings, but I don't know the output costs, will need to test 2023-02-23 02:42:17 +07:00
mrq 01c0941a40 binaries 2023-02-22 23:09:27 +07:00
mrq 4427d7fb84 initial conversion (errors out) 2023-02-22 23:07:05 +07:00
mrq 6c284ef8ec oops 2023-02-18 03:27:04 +07:00
mrq 8db762fa17 thought I copied this over 2023-02-18 03:15:44 +07:00
mrq 73d9c3bd46 set output folder to be sane with the cwd as a reference point 2023-02-18 02:01:09 +07:00
mrq 5ecf7da881 Fix later 2023-02-17 20:49:29 +07:00
mrq e3e8801e5f Fix I thought wasn't needed since it literally worked without it earlier 2023-02-17 20:41:20 +07:00
mrq 535549c3f3 add some snark about the kludge I had to fix, and the kludge I used to fix it 2023-02-17 19:20:19 +07:00
mrq a09cf98c7f more cleanup, pip-ifying won't work, got an alternative 2023-02-17 15:47:55 +07:00
mrq 6afa2c299e break if your dataset size is smaller than your batch size 2023-02-17 04:08:27 +07:00
mrq 94d0f16608 Necessary fixes to get it to work 2023-02-17 02:03:00 +07:00
mrq 49e23b226b pip-ify 2023-02-17 00:33:50 +07:00
James Betker f31a333c4f more sampling fixes 2022-10-10 20:11:28 +07:00
James Betker 5d172fbf7e Fix eval 2022-10-10 14:22:36 +07:00
James Betker 9502e0755e ugh 2022-10-10 12:15:51 +07:00
James Betker fce2c8f5db and listify them 2022-10-10 12:13:49 +07:00
James Betker 3cf78e3c44 train mel head even when not 2022-10-10 12:10:56 +07:00
James Betker cc74a43675 Checkin 2022-10-10 11:30:20 +07:00
James Betker 3cb14123bc glc fix 2022-07-29 11:24:36 +07:00
James Betker 4ddd01a7fb support generating cheaters from the new cheater network 2022-07-29 09:19:20 +07:00
James Betker 27a9b1b750 rename perplexity->log perplexity 2022-07-28 09:48:40 +07:00
James Betker 1d68624828 fix some imports.. 2022-07-28 02:35:32 +07:00
James Betker cfe907f13f i like this better 2022-07-28 02:33:23 +07:00
James Betker d44ed5d12d probably too harsh on ninfs 2022-07-28 01:33:54 +07:00
James Betker 4509cfc705 track logperp for diffusion evals 2022-07-28 01:30:44 +07:00
James Betker 19eb939ccf gd perplexity
# Conflicts:
#	codes/trainer/eval/music_diffusion_fid.py
2022-07-28 00:25:05 +07:00
James Betker a1bbde8a43 few things 2022-07-26 11:52:03 +07:00
James Betker f8108cfdb2 update environment and fix a bunch of deps 2022-07-24 23:43:25 +07:00
James Betker 45afefabed fix booboo 2022-07-24 18:00:14 +07:00
James Betker cc62ba9cba few more tfd13 things 2022-07-24 17:39:33 +07:00
James Betker f3d967dbf5 remove eta from mdf 2022-07-24 17:21:20 +07:00
James Betker 76464ca063 some fixes to mdf to support new archs 2022-07-21 10:55:50 +07:00
James Betker 13c263e9fb go all in on m2wv3 2022-07-21 00:51:27 +07:00
James Betker 24a78bd7d1 update tfd14 too 2022-07-21 00:45:33 +07:00