74 Commits

Author	SHA1	Message	Date
Alex O'Connell	9a6cde3684	detect experimental models in integration	2024-06-15 20:19:23 -04:00
Alex O'Connell	7393eae33d	Release v0.3.3	2024-06-15 18:18:42 -04:00
Alex O'Connell	1c3e532af4	improve generic openai backend	2024-06-11 21:21:46 -04:00
Alex O'Connell	35b766d540	Release v0.3.2	2024-06-08 16:38:17 -04:00
Alex O'Connell	0a6e558fda	Release v0.3.1	2024-06-08 12:14:24 -04:00
Alex O'Connell	249298bb99	better device prompting with area support + fix circular import	2024-06-07 17:42:03 -04:00
Alex O'Connell	5ddf0d09d5	restrict services that can be called and add format url function to make behavior standard	2024-06-07 08:48:50 -04:00
Alex O'Connell	21640dc321	release notes, fix service call args, and other release prep/cleanup	2024-06-06 23:54:08 -04:00
Alex O'Connell	ab32942006	more fixes for llm API	2024-06-06 23:08:04 -04:00
Alex O'Connell	36e29bedf0	add an LLM API to support the existing models	2024-06-06 22:40:59 -04:00
Alex O'Connell	b50904d73b	handle multi-turn tool models	2024-06-02 22:13:54 -04:00
Alex O'Connell	dbed6de6cd	Finish renaming stuff	2024-06-02 13:02:57 -04:00
Alex O'Connell	00d002d9c0	make tool formats work + dynamic quantization detection from HF	2024-06-02 12:25:26 -04:00
Alex O'Connell	8546767310	version with working ICL using the new APIs	2024-06-01 23:06:42 -04:00
Alex O'Connell	367607b14f	more rewrite work for new LLM API	2024-05-25 21:24:45 -04:00
Alex O'Connell	105f09ba2c	Release v0.2.17	2024-05-08 21:00:06 -04:00
Alex O'Connell	179e794283	add cmdline arguments to translate script + add defaults for command r	2024-05-08 20:50:25 -04:00
Stefan Daniel Schwarz	7404b6b36c	command-r prompt template (#141)	2024-05-07 22:33:47 +00:00
Alex O'Connell	6f2ce8828e	Release v0.2.16	2024-05-04 13:28:51 -04:00
Alex O'Connell	026a95d576	add prompt format for phi-3	2024-05-04 12:59:48 -04:00
Alex O'Connell	9eacd3edb2	Release v0.2.15	2024-05-04 07:33:39 -04:00
Alex O'Connell	cdd7e8415a	hook up flash attention	2024-05-02 23:05:43 -04:00
Alex O'Connell	875547d2e2	Release v0.2.14	2024-05-02 21:53:14 -04:00
Alex O'Connell	8d6116d4c3	bump version numbers	2024-04-24 21:25:27 -04:00
Alex O'Connell	92058d0694	manually set model language, zephyr2 prompt format , upgrading llama-cpp-python	2024-04-24 21:00:54 -04:00
Alex O'Connell	573207f7ff	explicit model match + fix llama 3	2024-04-21 23:42:32 -04:00
Alex O'Connell	adae87addd	training fixes, default values + other fixes	2024-04-21 23:40:28 -04:00
Alex O'Connell	3326bd7d6e	handle other languages in component	2024-04-21 20:38:46 -04:00
Alex O'Connell	d61b9b9242	wrong version number	2024-04-11 00:23:13 -04:00
Alex O'Connell	7262a2057a	add min p and typical p samplers	2024-04-10 23:55:01 -04:00
Alex O'Connell	1b22c06215	finish agent tests	2024-04-10 23:22:22 -04:00
Alex O'Connell	1577950137	more tests, fix missing default options, and load ICL as utf8	2024-04-08 20:58:07 -04:00
Alex O'Connell	9c3a3db696	more random settings	2024-04-07 22:27:13 -04:00
Alex O'Connell	5def7669f0	add recommended models	2024-04-06 22:55:52 -04:00
Alex O'Connell	5afc5014eb	Release v0.2.11	2024-04-06 20:13:01 -04:00
Alex O'Connell	793b36f215	Merge branch 'feature/prime-kv-cache' into develop	2024-04-06 18:24:18 -04:00
Alex O'Connell	f226dda4fc	Use github actions to build wheels	2024-04-06 18:03:53 -04:00
Alex O'Connell	4058a42ee5	clean up UI + hook up other llama.cpp settings	2024-03-31 17:08:52 -04:00
Alex O'Connell	fdd9f1bc67	wire up llama cpp runtime options	2024-03-31 11:08:07 -04:00
Alex O'Connell	0008f6bb5e	hook up prompt caching to config flow	2024-03-30 16:50:24 -04:00
Alex O'Connell	bf04cc3e6e	remove default host since it didn't make sense	2024-03-24 11:02:49 -04:00
Alex O'Connell	46e1c4fc1d	reset model_config when starting the configflow	2024-03-24 10:56:50 -04:00
Alex O'Connell	ac7b71ca4f	Make initial configuration easier + rewrite quickstart guide	2024-03-24 00:05:07 -04:00
Alex O'Connell	fa31682c51	working version of in context examples	2024-03-20 23:03:31 -04:00
Alex O'Connell	4978901412	start working on icl examples	2024-03-07 18:14:44 -05:00
Alex O'Connell	4f6ed08be9	split out service call argument allow list + properly parse rgb color arguments	2024-03-06 17:58:14 -05:00
Alex O'Connell	1c5414b8af	fix ollama keep alive properly + check if model exists for ollama too	2024-03-04 22:40:22 -05:00
Alex O'Connell	ff13770f7e	add Zephyr prompt format	2024-02-08 20:40:05 -05:00
Alex O'Connell	cc48465575	Add support for HTTPS endpoints	2024-02-08 19:26:58 -05:00
Isabella Nightshade	b1fb7cf184	Config to remember conversation/limit number of messages (#53)	2024-02-08 01:21:28 +00:00