2 Commits

Author SHA1 Message Date
c78ca0933c Update Helm release ingress-nginx to v4.13.3 2025-10-01 00:00:30 +00:00
708ffe203c Add Qwen2.5-VL models 2025-09-13 02:42:21 +02:00
3 changed files with 44 additions and 2 deletions

View File

@@ -54,7 +54,10 @@ spec:
"gemma3-4b", "gemma3-4b",
"gemma3-4b-novision", "gemma3-4b-novision",
"Qwen3-4B-Thinking-2507", "Qwen3-4B-Thinking-2507",
"Qwen3-4B-Thinking-2507-long-ctx" "Qwen3-4B-Thinking-2507-long-ctx",
"Qwen2.5-VL-7B-Instruct-GGUF",
"Qwen2.5-VL-32B-Instruct-GGUF-IQ1_S",
"Qwen2.5-VL-32B-Instruct-GGUF-Q2_K_L"
] ]
titleConvo: true titleConvo: true
titleModel: "gemma3-4b-novision" titleModel: "gemma3-4b-novision"

View File

@@ -175,3 +175,42 @@ models:
--flash-attn --flash-attn
--cache-type-k q8_0 --cache-type-v q8_0 --cache-type-k q8_0 --cache-type-v q8_0
--port ${PORT} --port ${PORT}
"Qwen2.5-VL-32B-Instruct-GGUF-IQ1_S":
ttl: 600
cmd: |
/app/llama-server
-hf unsloth/Qwen2.5-VL-32B-Instruct-GGUF:IQ1_S
-ngl 99 -c 16384 --predict 8192
--temp 0.7
--min-p 0.00
--top-p 0.8
--top-k 20
--repeat-penalty 1.0
--no-warmup
--port ${PORT}
"Qwen2.5-VL-32B-Instruct-GGUF-Q2_K_L":
ttl: 600
cmd: |
/app/llama-server
-hf unsloth/Qwen2.5-VL-32B-Instruct-GGUF:Q2_K_L
-ngl 99 -c 16384 --predict 8192
--temp 0.7
--min-p 0.00
--top-p 0.8
--top-k 20
--repeat-penalty 1.0
--no-warmup
--port ${PORT}
"Qwen2.5-VL-7B-Instruct-GGUF":
ttl: 600
cmd: |
/app/llama-server
-hf unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q4_K_M
-ngl 37 -c 16384 --predict 8192
--temp 0.7
--min-p 0.00
--top-p 0.8
--top-k 20
--repeat-penalty 1.0
--no-warmup
--port ${PORT}

View File

@@ -23,7 +23,7 @@ spec:
chart: chart:
spec: spec:
chart: ingress-nginx chart: ingress-nginx
version: 4.13.2 version: 4.13.3
sourceRef: sourceRef:
kind: HelmRepository kind: HelmRepository
name: ingress-nginx name: ingress-nginx