Commit Graph
405 Commits
Author SHA1 Message Date
me 7e2dd72fc1 llm, enable checking qwen port. 2026-10-12 09:52:19 -07:00
me b3a11dec30 llm, add qwen-code as a daemon, it seems promising, what I was hoping for from zeroclaw. 2026-10-12 09:43:15 -07:00
me 883ad223c2 llm, further simplify the pull-llm function so that it never exits prematurely or kills the terminal session. 2026-10-12 08:40:07 -07:00
me 9dee72820b llm, print the pull command in case the command keeps failing and we want to just do it manually. 2026-10-12 08:15:14 -07:00
me a3ba5e7a60 llm, try doing pull-models a different way since it's been trhowing weird errors. 2026-10-12 08:13:40 -07:00
me d27d759c97 llm, add TBD's for qwen serve, seems like it may be a really useful interface after a little playing, what I was hoping zeroclaw could do. 2026-10-11 06:04:13 -07:00
me a0eae9a7cc llm, make the context consistent across devices. 2026-10-11 05:06:54 -07:00
me df0df85d9e llm, use a more distinct number for the context so that ollama ps better gives away how the model is being used. 2026-10-10 19:46:20 -07:00
me 5541c8251c qwen3-coder is actually pretty great to chat with in ollama run, very responsive unlike in all the other tools. 2026-10-10 19:21:02 -07:00
me 487f1f46f3 llm, add Swap Stopper as a service based on custom shell script. 2026-10-10 18:41:50 -07:00
me ccd8355946 llm, fix ollama-stop function. 2026-10-10 17:27:47 -07:00
me 019066c4dc llm, further reduce contexts based on testing, now using 2.5:14b for Complete successfully alongside the RAM needs of 3:30b on Cline. 2026-10-10 16:42:19 -07:00
me 1f16053042 llm, continue tuning the model context and tokens to prevent swapping. 2026-10-10 16:04:09 -07:00
me 5d352bbb53 llm, add section comments, ensure ollama allows Codium workflows with concurrency and parallel. 2026-10-10 14:32:28 -07:00
me 6c46439eb6 llm, add shortcuts for stopping all ollama models. 2026-10-10 13:42:36 -07:00
me 648501b7c0 llm, use pipefail on the final lint error. 2026-10-10 13:35:34 -07:00
me e0288ce4d5 llm, filter models by purpose, use chat in Zeroclaw and code in Continue. 2026-10-10 13:33:48 -07:00
me c6cc631d8c llm, tuning the context and token parameters. 2026-10-10 13:10:29 -07:00
me 53351fde28 llm, add a finish for open-webui, fix config.yaml for Continue, possibly fix pull-models causing terminals to close when cancelled. 2026-10-10 11:12:25 -07:00
me 7922225b26 Remove unecessary update_cache which eat up time. 2026-10-10 11:11:15 -07:00
me 8970b14e31 llm, always do the setup steps, then XYZ_enabled vars are used to disable and stop services not wanted. 2026-10-10 10:28:45 -07:00
me c85d96f940 llm, add hugging face token to ollama. 2026-10-10 09:56:41 -07:00
me 4ab93d02f4 llm, fix the hugging face token env var for open-webui. 2026-10-10 09:49:29 -07:00
me 0feb9b1995 llm, fix filename for continue, add print for llm check, refer to stdout on llm check. 2026-10-10 04:57:00 -07:00
me e95054d329 llm, use new variables provisioned for open-webui. 2026-10-10 04:51:10 -07:00
me daec9237f0 llm, finish the open-webui setup, start using the workstation config variables. 2026-10-10 04:20:28 -07:00
me 196c879c27 llm, begin adding section for open-webui, untested so far. 2026-10-09 20:23:17 -07:00
me 6735e257b3 llm, enhance comments in the ollama run script. 2026-10-09 20:06:14 -07:00
me 4ce0720a43 llm, change exit to return to prevent session from closing. 2026-10-09 20:02:06 -07:00
me c59d78270b llm, stop purging partial pulls. 2026-10-09 20:00:28 -07:00
me c7432d554f llm, add each LLM as an autocomplete agent for Continue. 2026-10-09 16:56:15 -07:00
me bd181de2c9 vpn, fix lint issues besides shell changed_when. 2026-10-09 16:47:59 -07:00
me e95c87dddf node, fix lint issues: fix indentation, do not update cache or use the latest file, add a changed_chen to shell. 2026-10-09 16:47:00 -07:00
me b5c7bed4e9 llm, change message to not display \n's. 2026-10-09 15:45:36 -07:00
me f6dfa4f81f llm, fix unfinished message. 2026-10-09 15:44:39 -07:00
me 9e6e162b03 llm, attempt to fix decimal values which are supposed to be integers, fix some spacing. 2026-10-09 15:26:22 -07:00
me 816ef4ddc6 llm, simplify Continue config, add a function for pulling the LLMs outside of Ansible, adjust ollama settings. 2026-10-09 15:18:03 -07:00
me 731a367e6f llm, add config for Continue (a Codium plugin), fix size_mb not actually getting calculated, change log files to be per-run by making the timestamp very specific, further adjust the ollama environment variables. 2026-10-09 14:52:27 -07:00
me ee97598070 llm, refactor zeroclaw.yml to be llm since zeroclaw is now understood to be less of the priority. 2026-10-09 13:19:34 -07:00
me 74b42f3d63 zeroclaw, variablize the context amount to be per-machine. 2026-10-09 13:18:07 -07:00
me 0c86a76968 zeroclaw, greatly simplify the choosing of LLMs based on resources, refactor ollama run file, modify some of the env values, add qwen-code, change some of the models in the download queue. 2026-10-09 11:43:46 -07:00
me a9080a6d12 zeroclaw, a few more small edits while learning how things work. 2026-10-08 21:09:24 -07:00
me 22252b8880 zeroclaw, refactor a bit more, add env variables to zeroclaw daemon. 2026-10-08 20:00:34 -07:00
me 9860f56bf7 zeroclaw, don't download qwen3.5:27b since 35b is also being used. 2026-10-08 12:09:02 -07:00
me 10821d2317 zeroclaw, still trying to fine-tune things, getting closer. 2026-10-08 10:38:05 -07:00
me 77d2ae62a3 zeroclaw, begin formatting the removals, add some more TBD's. 2026-10-07 11:58:16 -07:00
me a511fdac35 zeroclaw, move providers back after the installs so that they are accurate in the GUI. 2026-10-07 11:33:51 -07:00
me 934bec45b5 zeroclaw, remove extra newline. 2026-10-07 11:25:01 -07:00
me 0660dd55d4 zeroclaw, have both backup, not just the first one, in case the first doesn't change anything and that's what causes the save. 2026-10-07 11:20:39 -07:00
me 57707ff56e zeroclaw, add explicit attributes to the blockinfile tasks. 2026-10-07 11:19:50 -07:00