apunkt/trustgraph
1
0
Fork 0
mirror of https://github.com/trustgraph-ai/trustgraph.git synced 2026-04-26 00:46:22 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Dual model templates (#263)

Browse source 
* Dual-mode templates

* Fixed generate-all so that works

* Fix ability to specify Claude model
This commit is contained in:
cybermaggedon 2025-01-11 11:40:42 +00:00 • committed by GitHub
parent 1bb1112569
commit cec9e29222
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
24 changed files with 725 additions and 360 deletions
Download patch file Download diff file Expand all files Collapse all files

33
templates/components/llamafile.jsonnet
View file

@ -5,6 +5,11 @@ local prompts = import "prompts/slm.jsonnet";
{
with:: function(key, value)
self + {
["llamafile-" + key]:: value,
},
"llamafile-model":: "LLaMA_CPP",
"text-completion" +: {
@ -28,46 +33,18 @@ local prompts = import "prompts/slm.jsonnet";
.with_limits("0.5", "128M")
.with_reservations("0.1", "128M");
local containerRag =
engine.container("text-completion-rag")
.with_image(images.trustgraph)
.with_command([
"text-completion-llamafile",
"-p",
url.pulsar,
"-m",
$["llamafile-model"],
"-i",
"non-persistent://tg/request/text-completion-rag",
"-o",
"non-persistent://tg/response/text-completion-rag",
])
.with_env_var_secrets(envSecrets)
.with_limits("0.5", "128M")
.with_reservations("0.1", "128M");
local containerSet = engine.containers(
"text-completion", [ container ]
);
local containerSetRag = engine.containers(
"text-completion-rag", [ containerRag ]
);
local service =
engine.internalService(containerSet)
.with_port(8080, 8080, "metrics");
local serviceRag =
engine.internalService(containerSetRag)
.with_port(8080, 8080, "metrics");
engine.resources([
envSecrets,
containerSet,
containerSetRag,
service,
serviceRag,
])
},