1697 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
louistiti	dcda888de6	fix(server): persona typo	2024-07-04 09:21:02 +08:00
louistiti	0c775ba2e5	feat: use VRAM as LLM unit requirements	2024-07-03 23:41:23 +08:00
louistiti	6867e9c6db	feat(server): VRAM helpers	2024-07-03 13:42:10 +08:00
louistiti	1334c2013c	feat(server): has GPU helper	2024-07-03 13:34:05 +08:00
louistiti	94885de71b	feat(server): get graphics compute API	2024-07-03 13:31:26 +08:00
louistiti	b86775846b	feat(server): get GPU device names	2024-07-03 13:24:53 +08:00
louistiti	4019bd877b	chore: better comments on LLM action matching	2024-07-03 09:11:25 +08:00
louistiti	7e483b9123	feat: add inspect:gpu npm script	2024-07-02 22:12:48 +08:00
louistiti	ee615e6ade	refactor(python tcp server): rename the RMS threshold setting for ASR	2024-07-01 22:53:34 +08:00
louistiti	b092947e2e	feat(web app): add headset tips for a better voice experience	2024-07-01 22:52:16 +08:00
louistiti	1d96655b01	fix(python tcp server): overflowed on ASR	2024-07-01 22:28:18 +08:00
louistiti	8ed7c78074	fix(web app): use correct config property for LLM warm up	2024-07-01 09:45:58 +08:00
louistiti	be3df774dc	feat(server): boost free RAM delta for LLM	2024-07-01 09:38:20 +08:00
louistiti	2c89041b3f	feat(server): upgrade node-llama-cpp to 3.0.0-beta.36	2024-07-01 09:03:02 +08:00
louistiti	e4277bfcf4	feat(web app): add more info data	2024-06-30 20:04:08 +08:00
louistiti	2e351d4218	feat(web app): add info	2024-06-30 19:58:26 +08:00
louistiti	e24dee3b07	feat(server): VRAM context size management	2024-06-30 19:15:24 +08:00
louistiti	8eeff3a01c	feat(python tcp server): map speech synthesis hardware device choice to settings	2024-06-30 16:54:59 +08:00
louistiti	a05f34422e	feat(server): upgrade node-llama-cpp to 3.0.0-beta.34	2024-06-30 16:47:57 +08:00
louistiti	74241b43fe	feat(python tcp server): run speech synthesis inference on CPU	2024-06-30 16:30:12 +08:00
louistiti	fae15c9f4d	feat(server): kill existing PyTorch thread from TCP server on start	2024-06-30 16:03:11 +08:00
louistiti	2c1b35144d	feat(server): use debug verbosity by default in LLM manager	2024-06-30 15:54:51 +08:00
louistiti	c25c52532f	feat(server): increase LLM threads number	2024-06-30 15:54:12 +08:00
louistiti	805c65b2b2	fix(web app): init state when shouldWarmUpLLM is not enabled	2024-06-30 15:53:49 +08:00
louistiti	0bb61e5301	Merge remote-tracking branch 'origin/develop' into develop	2024-06-29 23:29:52 +08:00
louistiti	d2038061e7	feat(server): disable onToken when LLM duties are warming up	2024-06-29 23:29:25 +08:00
Louis Grenard	7a42f4df93	feat(server): disable onToken when LLM duties are warming up	2024-06-29 23:28:19 +08:00
louistiti	ed9961c822	feat(server): sync LLM duties warmup with UI	2024-06-27 23:36:32 +08:00
louistiti	fb91551364	feat(server): warm up LLM duties when necessary (WIP)	2024-06-25 23:49:09 +08:00
louistiti	8dc725a697	refactor(scripts): differentiate PyTorch info log on macOS	2024-06-24 08:28:32 +08:00
louistiti	6c3abf30b0	fix(server): sometimes the action recognition LLM duty add whitespace to intent	2024-06-23 23:36:11 +08:00
louistiti	1314217e00	BREAKING: upgrade from Python 3.9 to Python 3.11	2024-06-23 23:29:09 +08:00
louistiti	871617649a	chore(bridge/python): upgrade cx_Freeze to 7.1.1	2024-06-23 22:42:42 +08:00
louistiti	85c3a40457	chore(python tcp server): upgrade cx_Freeze to 7.1.1	2024-06-23 22:41:46 +08:00
louistiti	0661e9e04f	refactor(scripts): only install PyTorch when the targeted setup is the TCP server	2024-06-23 22:39:39 +08:00
louistiti	6045eb5105	feat(server): speed up translation LLM duty	2024-06-23 22:27:12 +08:00
louistiti	72954a3557	feat(server): speed up summarization LLM duty	2024-06-23 22:13:36 +08:00
louistiti	cd65b84b7c	feat(server): init LLM duty on inference requested by skills	2024-06-23 18:47:58 +08:00
louistiti	c0e9fdcd8b	feat(server): speed up paraphrase LLM duty	2024-06-23 18:46:32 +08:00
louistiti	780fd97006	feat(server): speed up NER custom LLM duty	2024-06-23 18:31:18 +08:00
louistiti	e985d65165	feat(server): improve action recognition LLM duty to not hallucinate	2024-06-23 17:45:29 +08:00
louistiti	223cfdc0c9	fix(server): force lowercase on intent name for the action recognition LLM duty	2024-06-23 17:43:29 +08:00
louistiti	38a700feb0	fix: whitelist config.json for the TTS model	2024-06-23 17:24:38 +08:00
louistiti	faee688924	feat: keep .gitkeep on audio models	2024-06-23 17:07:45 +08:00
louistiti	a9e5a2978f	feat(server): prepare LLM duties speed improvement on context/session creation	2024-06-22 23:23:39 +08:00
louistiti	ab0e1493a0	feat(server): speed up LLM action recognition	2024-06-22 23:19:21 +08:00
louistiti	76b197baeb	refactor(python tcp server): ASR CUDA log notification on bootup	2024-06-22 21:03:27 +08:00
louistiti	aa63a6dabc	chore(server): upgrade nodemon to 3.1.4	2024-06-22 12:44:23 +08:00
louistiti	1b4a1f15a9	feat(server): upgrade node-llama-cpp to 3.0.0-beta.32	2024-06-21 23:34:54 +08:00
louistiti	3734a16b6a	fix(python tcp server): TTS inference correct params mapping	2024-06-21 09:51:52 +08:00