314 commits

Author	SHA1	Message	Date
oobabooga	364529d0c7	Further refactor	2023-02-23 14:31:28 -03:00
oobabooga	e46c43afa6	Move some stuff from server.py to modules	2023-02-23 13:42:23 -03:00
oobabooga	1dacd34165	Further refactor	2023-02-23 13:28:30 -03:00
oobabooga	ce7feb3641	Further refactor	2023-02-23 13:03:52 -03:00
oobabooga	98af4bfb0d	Refactor the code to make it more modular	2023-02-23 12:05:25 -03:00
oobabooga	18e0ec955e	Improve some descriptions in --help	2023-02-23 10:11:58 -03:00
oobabooga	c72892835a	Don't show *-np models in the list of choices	2023-02-22 11:38:16 -03:00
oobabooga	044b963987	Add stop parameter for flexgen (#105)	2023-02-22 11:23:36 -03:00
oobabooga	ea21a22940	Remove redundant preset	2023-02-22 01:01:26 -03:00
oobabooga	b8b3d4139c	Add --compress-weight parameter	2023-02-22 00:43:21 -03:00
oobabooga	eef6fc3cbf	Add a preset for FlexGen	2023-02-21 23:33:15 -03:00
oobabooga	311404e258	Reuse disk-cache-dir parameter for flexgen	2023-02-21 22:11:05 -03:00
oobabooga	f3c75bbd64	Add --percent flag for flexgen	2023-02-21 22:08:46 -03:00
oobabooga	b83f51ee04	Add FlexGen support #92 (experimental)	2023-02-21 21:00:06 -03:00
oobabooga	444cd69c67	Fix regex bug in loading character jsons with special characters	2023-02-20 19:38:19 -03:00
oobabooga	d7a738fb7a	Load any 13b/20b/30b model in 8-bit mode when no flags are supplied	2023-02-20 15:44:10 -03:00
oobabooga	77846ceef3	Minor change	2023-02-20 15:05:48 -03:00
oobabooga	e195377050	Deprecate torch dumps, move to safetensors (they load even faster)	2023-02-20 15:03:19 -03:00
oobabooga	14ffa0b418	Fix line breaks in --chat mode	2023-02-20 13:25:46 -03:00
SillyLossy	ded890c378	Escape regexp in message extraction	2023-02-19 12:55:45 +02:00
oobabooga	8c9dd95d55	Print the softprompt metadata when it is loaded	2023-02-19 01:48:23 -03:00
oobabooga	f79805f4a4	Change a comment	2023-02-18 22:58:40 -03:00
oobabooga	d58544a420	Some minor formatting changes	2023-02-18 11:07:55 -03:00
oobabooga	0dd41e4830	Reorganize the sliders some more	2023-02-17 16:33:27 -03:00
oobabooga	6b9ac2f88e	Reorganize the generation parameters	2023-02-17 16:18:01 -03:00
oobabooga	596732a981	The soft prompt length must be considered here too	2023-02-17 12:35:30 -03:00
oobabooga	edc0262889	Minor file uploading fixes	2023-02-17 10:27:41 -03:00
oobabooga	243244eeec	Attempt at fixing greyed out files on iphone	2023-02-17 10:17:15 -03:00
oobabooga	a226f4cddb	No change, so reverting	2023-02-17 09:27:17 -03:00
oobabooga	40cb9f63f6	Try making Colab happy (tensorflow warnings)	2023-02-17 09:23:11 -03:00
oobabooga	aeddf902ec	Make the refresh button prettier	2023-02-16 21:55:20 -03:00
oobabooga	21512e2790	Make the Stop button work more reliably	2023-02-16 21:21:45 -03:00
oobabooga	08805b3374	Force "You" in impersonate too	2023-02-16 13:24:13 -03:00
oobabooga	d7db04403f	Fix --chat chatbox height	2023-02-16 12:45:05 -03:00
oobabooga	589069e105	Don't regenerate if no message has been sent	2023-02-16 12:32:35 -03:00
oobabooga	405dfbf57c	Force your name to be "You" for pygmalion (properly)	2023-02-16 12:16:12 -03:00
oobabooga	7bd2ae05bf	Force your name to be "You" for pygmalion ... This allows you to customize your displayed name.	2023-02-15 21:32:53 -03:00
oobabooga	3746d72853	More style fixes	2023-02-15 21:13:12 -03:00
oobabooga	6f213b8c14	Style fix	2023-02-15 20:58:17 -03:00
oobabooga	ccf10db60f	Move stuff into tabs in chat mode	2023-02-15 20:55:32 -03:00
oobabooga	a55e8836f6	Bump gradio version ... It looks uglier, but the old one was bugged and unstable.	2023-02-15 20:20:56 -03:00
oobabooga	0e89ff4b13	Clear the persistent history after clicking on "Clear history"	2023-02-15 16:49:52 -03:00
oobabooga	b3bcd2881d	Implement regenerate/impersonate the proper way (fixes #78)	2023-02-15 14:39:26 -03:00
oobabooga	5ee9283cae	Mention BLIP	2023-02-15 13:53:38 -03:00
oobabooga	8d3b3959e7	Document --picture option	2023-02-15 13:50:18 -03:00
oobabooga	2eea0f4edb	Minor change	2023-02-15 12:58:11 -03:00
oobabooga	3c31fa7079	Simplifications	2023-02-15 12:46:11 -03:00
oobabooga	80fbc584f7	Readability	2023-02-15 11:38:44 -03:00
oobabooga	b397bea387	Make chat history persistent	2023-02-15 11:30:38 -03:00
oobabooga	7be372829d	Set chat prompt size in tokens	2023-02-15 10:18:50 -03:00
oobabooga	8c3ef58e00	Use BLIP directly + some simplifications	2023-02-14 23:55:46 -03:00
SillyLossy	a7d98f494a	Use BLIP to send a picture to model	2023-02-15 01:38:21 +02:00
oobabooga	d910d435cd	Consider the softprompt in the maximum prompt length calculation	2023-02-14 12:06:47 -03:00
oobabooga	8b3bb512ef	Minor bug fix (soft prompt was being loaded twice)	2023-02-13 23:34:04 -03:00
oobabooga	7739a29524	Some simplifications	2023-02-13 18:48:32 -03:00
oobabooga	3277b751f5	Add softprompt support (for real this time) ... Is this too much voodoo for our purposes?	2023-02-13 15:25:16 -03:00
oobabooga	aa1177ff15	Send last internal reply to input rather than visible	2023-02-13 03:29:23 -03:00
oobabooga	2c3abcf57a	Add support for rosey/chip/joi instruct models	2023-02-12 09:46:34 -03:00
oobabooga	7ef7bba6e6	Add progress bar for model loading	2023-02-12 09:36:27 -03:00
oobabooga	5d3f15b915	Use the CPU if no GPU is detected	2023-02-11 23:17:06 -03:00
oobabooga	b3c4657c47	Remove commas from preset files	2023-02-11 14:54:29 -03:00
oobabooga	0dd1409f24	Add penalty_alpha parameter (contrastive search)	2023-02-11 14:48:12 -03:00
oobabooga	2ed0386d87	Fix replace last reply in --chat mode (for #69)	2023-02-11 07:59:54 -03:00
oobabooga	316e07f06a	auto-assign gpu memory with --auto-devices alone	2023-02-10 16:36:06 -03:00
oobabooga	219366342b	Sort imports according to PEP8 (based on #67)	2023-02-10 15:40:03 -03:00
81300	20dbef9623	Extend bfloat16 support	2023-02-09 20:00:03 +02:00
oobabooga	cadd100405	min_length has to be 0 when streaming is on	2023-02-08 00:23:35 -03:00
oobabooga	6be571cff7	Better variable names	2023-02-08 00:19:20 -03:00
oobabooga	58b07cca81	length_penalty can be negative (apparently)	2023-02-07 23:33:02 -03:00
oobabooga	7e4c25691d	Repetition penalty has to be < 5	2023-02-07 23:23:39 -03:00
oobabooga	1c30e1b49a	Add even more sliders	2023-02-07 23:11:04 -03:00
oobabooga	24dc705eca	Add lots of sliders	2023-02-07 22:08:21 -03:00
Martin J	06a4664805	Fix a regex issue in tokenize_dialogue. ... The existing regex would fail if using character names that start with numbers, for example: 9S or 2B.	2023-02-05 07:42:57 +01:00
oobabooga	2fe235738e	Reorganize chat buttons	2023-02-04 22:53:42 -03:00
oobabooga	2207d44986	Windows doesn't like : in filenames	2023-02-04 20:07:39 -03:00
oobabooga	65266f3349	Fix loading official colab chat logs	2023-02-03 22:43:02 -03:00
oobabooga	44e8c671f9	Fix API documentation formatting in chat mode	2023-02-03 10:00:05 -03:00
oobabooga	a28f0d8bd7	Show it/s in the same units with or without streaming ... Closes #49	2023-02-03 09:11:11 -03:00
oobabooga	4e4cd67223	Save chat history with name/date in filename ... closes #50	2023-02-03 09:02:35 -03:00
oobabooga	3af3ffeb90	Make --help output more readable	2023-02-02 23:36:28 -03:00
oobabooga	638495b633	Simplify generate() function	2023-02-02 13:47:08 -03:00
oobabooga	3f05cf5ddd	Simplify encode() function	2023-02-02 13:31:32 -03:00
oobabooga	2583bc5840	Simplify deepspeed implementation (#40)	2023-02-02 12:15:44 -03:00
oobabooga	f38c9bf428	Fix deepspeed (oops)	2023-02-02 10:39:37 -03:00
oobabooga	90f1067598	Move deepspeed parameters to another file	2023-02-02 10:25:09 -03:00
81300	248ec4fa21	Merge branch 'oobabooga:main' into ds	2023-02-01 20:50:51 +02:00
81300	a6f4760772	Add arg for bfloat16	2023-02-01 20:22:07 +02:00
81300	c515282f5c	no_split_module_classes not needed	2023-02-01 19:47:26 +02:00
81300	0a0d289537	Fix issue with generating on multiple GPUs	2023-02-01 19:02:07 +02:00
81300	a97afa6965	Add DeepSpeed ZeRO-3 integration	2023-02-01 18:48:13 +02:00
oobabooga	6b13816c47	Change default --disk behavior	2023-02-01 10:43:28 -03:00
oobabooga	119be56390	Add back low_cpu_mem_usage=True ... Removing it didn't help with anything, so I am adding it bad on a purely superstiticious basis.	2023-02-01 10:01:44 -03:00
oobabooga	d4a0b377ab	Allow standalone --cpu-memory ... I think that what I am doing probably makes sense, but I could be wrong.	2023-01-31 21:23:16 -03:00
oobabooga	8ef89df746	Try to leave at least 1GiB free to prevent oom errors	2023-01-31 20:47:05 -03:00
oobabooga	bb77f20a6c	Don't use low_cpu_mem_usage and device_map together	2023-01-31 13:24:05 -03:00
oobabooga	001ecf95b2	Update server.py	2023-01-31 08:14:16 -03:00
Silver267	a85bb5e9a2	Fix an error ... Fixes "UnboundLocalError: local variable 'substring_found' referenced before assignment" when loading non-pygmalion models in cai chat mode.	2023-01-31 01:34:10 -05:00
oobabooga	5b0bbfa6e8	Clean up	2023-01-30 14:17:12 -03:00
oobabooga	2dadf42cb5	Print the tokenized example dialogue in a prettier way	2023-01-30 08:29:49 -03:00
oobabooga	161cae001b	I needed this	2023-01-29 23:20:22 -03:00