simcop2387/llama-swap-infra
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages 8 Wiki Activity Actions
No description
872 commits 3 branches 13 tags 879 KiB
  • Jinja 61.2%
  • Dockerfile 26.2%
  • Perl 11.4%
  • Shell 1.2%
Find a file
Exact
Git bot 4cdea2f01e
Some checks failed
/ build-vllm (push) Successful in 23m11s
Details
/ build-swap (push) Successful in 19m21s
Details
/ build-comfyui (push) Successful in 27m20s
Details
/ build-llama-server (push) Successful in 1h28m11s
Details
/ build-ik-llama-server (push) Successful in 2h35m4s
Details
/ build-combined (push) Failing after 26m31s
Details
/ Clean up stuff (push) Successful in 2s
Details
Auto updated submodule references
2026-05-11 05:31:41 +00:00
.forgejo/workflows wtf is going on here? 2026-05-08 17:10:44 -04:00
ComfyUI@52976f3ea3 Auto updated submodule references 2026-05-11 05:31:41 +00:00
config-gen Generates what looks like a working config now 2025-05-24 19:20:57 -04:00
ik_llama.cpp@23127139cb Auto updated submodule references 2026-05-11 05:31:41 +00:00
litellm-llama-swap-terraform@1ca955f4c2 Add the slot overrides for the new cpu models 2026-05-08 11:42:54 -04:00
llama-swap@2be3416baa Auto updated submodule references 2026-05-10 05:31:47 +00:00
llama.cpp@7d442abf5c Auto updated submodule references 2026-05-11 05:31:41 +00:00
prompt-templates Attempt at this with chatml template bits 2026-02-04 15:35:40 -05:00
swap-configs Drop the qwen3-tts server thats run this way. makes for annoying setups 2026-05-08 11:22:03 -04:00
vllm-mirror@bcc6f67447 upgrade vllm bits to 0.18.X for jina embed v5 2026-04-01 09:06:44 -04:00
.gitmodules try swapping to https? 2026-03-29 10:58:24 -04:00
.yamllint.yml Fixup some linting stuff and continue adding comfyui 2026-02-21 09:16:00 -05:00
AGENTS.md chore: add agents configuration documentation 2026-01-16 09:10:54 -05:00
base-image.Containerfile mostly rework the python stuff, not building proper combined image yet 2026-03-03 17:51:01 -05:00
combined.Containerfile ugh wrong one 2026-03-12 11:21:54 -04:00
comfyui-image.Containerfile Redo the fucking comfyui bits because theyve changed stuff that makes it impossible to build the old way. need to investigate how to otherwise build it from my local copy still 2026-03-30 09:55:42 -04:00
compose-base.yml Try switching... 2026-03-30 19:08:30 -04:00
compose-brainiac.yml Add container for cpu runner 2026-05-08 11:10:08 -04:00
compose-manchester.yml update manchester a bit again 2026-03-31 11:09:01 -04:00
data.json small changes, see how this helps 2025-08-13 10:39:41 -04:00
ik-llama-server.Containerfile fix accidental comfy dep 2026-03-06 18:47:00 -05:00
llama-server.Containerfile fix accidental comfy dep 2026-03-06 18:47:00 -05:00
llama-swap.Containerfile separated base image piece 2025-11-30 13:29:37 -05:00
notes_for_comfyui-qwen3-tts.txt Temporary notes and changes for qwen3tts comfy 2026-01-26 19:26:52 -05:00
qwen3-tts-server.sh Doh! use exec to make unloading work more reliably! 2026-01-26 12:45:42 -05:00
README.md add readme to see what happens with actions 2025-08-14 14:03:26 -04:00
run-comfy.sh mostly rework the python stuff, not building proper combined image yet 2026-03-03 17:51:01 -05:00
send_komodo_webhook.sh Add automatic deploy webhook also 2025-08-14 14:16:58 -04:00
vllm-image.Containerfile fix version here 2026-04-01 09:52:48 -04:00

README.md

A set of configs and container builds for llama-swap on my custom infrastructure. Soon to add sdome arm64 stuff for a jetson xavier but not ready yet for that.