Compare commits
61 Commits
b7d739bb6e
...
renovate/c
| Author | SHA1 | Date | |
|---|---|---|---|
| d0ef8f30b9 | |||
| 98f63b1576 | |||
| edba33b552 | |||
| 054df42d8b | |||
| 08db022d0d | |||
| e485a4fc7f | |||
| 9e74ed6a19 | |||
| 42e89c9bb7 | |||
| 99bc04b76a | |||
| 7ee77e33d4 | |||
| 8bdd5f2196 | |||
| 1d8cb85bd4 | |||
| eeb302b63b | |||
| 69b437ed3b | |||
| 54674a6e79 | |||
| a9da405326 | |||
| 264871bf68 | |||
| 6bcd0ba464 | |||
| cb53301926 | |||
| 110817b748 | |||
| 66cb3c9d82 | |||
| 42ae7af649 | |||
| cffcb1cc2d | |||
| a4a7dd6fe6 | |||
| 52b8ca79dc | |||
| 9a1fe1f740 | |||
| e996a60378 | |||
| 0ccd4d93f1 | |||
| d667c6c0fc | |||
| 4254ebc9ef | |||
| 8cf02fea0e | |||
| aa3c74d6a7 | |||
| 289089428e | |||
| a93f6ec36f | |||
| 1d85bf3a88 | |||
| f495debf25 | |||
| bfede17c87 | |||
| 08ca3f4c4e | |||
| 471c0ba62d | |||
| 261141f509 | |||
| 86d5751842 | |||
| 43e531a3ca | |||
| 9a0764268b | |||
| 7c88498756 | |||
| 8717526358 | |||
| b6a7e5092c | |||
| 27f7a5f29a | |||
| 9d0fd0981a | |||
| 51bc53dbbc | |||
| ce0b13ebb3 | |||
| 516e157d39 | |||
| 73d6d1f15a | |||
| c51fc2a5ef | |||
| 8d994e7aa1 | |||
| 5b551c6c6e | |||
| 7e7b3e3d71 | |||
| 9f315b38e3 | |||
| 3e1a806db1 | |||
| f7dba45165 | |||
| c8fac3201a | |||
| 82864a4738 |
@@ -15,8 +15,8 @@ spec:
|
||||
spec:
|
||||
containers:
|
||||
- name: crawl4ai-proxy
|
||||
image: ghcr.io/lennyerik/crawl4ai-proxy:latest
|
||||
imagePullPolicy: IfNotPresent
|
||||
image: gitea.lumpiasty.xyz/lumpiasty/crawl4ai-proxy-fit:latest
|
||||
imagePullPolicy: Always
|
||||
env:
|
||||
- name: LISTEN_PORT
|
||||
value: "8000"
|
||||
|
||||
@@ -18,7 +18,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: immich
|
||||
version: 1.1.1
|
||||
version: 1.2.2
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: secustor
|
||||
|
||||
@@ -4,10 +4,13 @@ logToStdout: "both" # proxy and upstream
|
||||
|
||||
macros:
|
||||
base_args: "--no-warmup --port ${PORT}"
|
||||
common_args: "--fit-target 1536 --fit-ctx 65536 --no-warmup --port ${PORT}"
|
||||
common_args: "--fit-target 1536 --no-warmup --port ${PORT}"
|
||||
gemma3_ctx_128k: "--ctx-size 131072"
|
||||
qwen35_ctx_128k: "--ctx-size 131072"
|
||||
qwen35_ctx_256k: "--ctx-size 262144"
|
||||
gemma_sampling: "--prio 2 --temp 1.0 --repeat-penalty 1.0 --min-p 0.00 --top-k 64 --top-p 0.95"
|
||||
qwen35_sampling: "--temp 0.6 --top-p 0.95 --top-k 20 --min-p 0.00 -ctk q4_0 -ctv q4_0"
|
||||
qwen35_35b_args: "--temp 1.0 --min-p 0.00 --top-p 0.95 --top-k 20 -ctk q4_0 -ctv q4_0"
|
||||
qwen35_sampling: "--temp 0.6 --top-p 0.95 --top-k 20 --min-p 0.00 -ctk q8_0 -ctv q8_0"
|
||||
qwen35_35b_args: "--temp 1.0 --min-p 0.00 --top-p 0.95 --top-k 20 -ctk q8_0 -ctv q8_0"
|
||||
qwen35_35b_heretic_mmproj: "--mmproj-url https://huggingface.co/unsloth/Qwen3.5-35B-A3B-GGUF/resolve/main/mmproj-F16.gguf --mmproj /root/.cache/llama.cpp/unsloth_Qwen3.5-35B-A3B-GGUF_mmproj-F16.gguf"
|
||||
qwen35_4b_heretic_mmproj: "--mmproj-url https://huggingface.co/unsloth/Qwen3.5-4B-GGUF/resolve/main/mmproj-F16.gguf --mmproj /root/.cache/llama.cpp/unsloth_Qwen3.5-4B-GGUF_mmproj-F16.gguf"
|
||||
glm47_flash_args: "--temp 0.7 --top-p 1.0 --min-p 0.01 --repeat-penalty 1.0"
|
||||
@@ -39,6 +42,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/gemma-3-12b-it-GGUF:Q4_K_M
|
||||
${gemma3_ctx_128k}
|
||||
${gemma_sampling}
|
||||
${common_args}
|
||||
|
||||
@@ -46,6 +50,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/gemma-3-12b-it-GGUF:Q4_K_M
|
||||
${gemma3_ctx_128k}
|
||||
${gemma_sampling}
|
||||
--no-mmproj
|
||||
${common_args}
|
||||
@@ -54,6 +59,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/gemma-3-4b-it-GGUF:Q4_K_M
|
||||
${gemma3_ctx_128k}
|
||||
${gemma_sampling}
|
||||
${common_args}
|
||||
|
||||
@@ -61,6 +67,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/gemma-3-4b-it-GGUF:Q4_K_M
|
||||
${gemma3_ctx_128k}
|
||||
${gemma_sampling}
|
||||
--no-mmproj
|
||||
${common_args}
|
||||
@@ -76,13 +83,14 @@ models:
|
||||
--top-p 0.95
|
||||
--top-k 40
|
||||
--repeat-penalty 1.0
|
||||
-ctk q4_0 -ctv q4_0
|
||||
-ctk q8_0 -ctv q8_0
|
||||
${common_args}
|
||||
|
||||
"Qwen3.5-35B-A3B-GGUF:Q4_K_M":
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-35B-A3B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_35b_args}
|
||||
${common_args}
|
||||
|
||||
@@ -90,6 +98,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-35B-A3B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_35b_args}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -101,6 +110,7 @@ models:
|
||||
/app/llama-server
|
||||
-hf mradermacher/Qwen3.5-35B-A3B-heretic-GGUF:Q4_K_M
|
||||
${qwen35_35b_heretic_mmproj}
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_35b_args}
|
||||
${common_args}
|
||||
|
||||
@@ -109,6 +119,7 @@ models:
|
||||
/app/llama-server
|
||||
-hf mradermacher/Qwen3.5-35B-A3B-heretic-GGUF:Q4_K_M
|
||||
${qwen35_35b_heretic_mmproj}
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_35b_args}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -117,6 +128,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-0.8B-GGUF:Q4_K_XL
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${base_args}
|
||||
${thinking_on}
|
||||
@@ -134,6 +146,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-2B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -142,6 +155,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-2B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -150,6 +164,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-4B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_128k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -158,6 +173,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-4B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_128k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -167,6 +183,7 @@ models:
|
||||
/app/llama-server
|
||||
-hf mradermacher/Qwen3.5-4B-heretic-GGUF:Q4_K_M
|
||||
${qwen35_4b_heretic_mmproj}
|
||||
${qwen35_ctx_128k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -176,6 +193,7 @@ models:
|
||||
/app/llama-server
|
||||
-hf mradermacher/Qwen3.5-4B-heretic-GGUF:Q4_K_M
|
||||
${qwen35_4b_heretic_mmproj}
|
||||
${qwen35_ctx_128k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -184,6 +202,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-9B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -192,6 +211,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-9B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -200,6 +220,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-9B-GGUF:Q3_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -208,6 +229,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-9B-GGUF:Q3_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -216,6 +238,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-27B-GGUF:Q3_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -224,6 +247,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-27B-GGUF:Q3_K_M
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
|
||||
@@ -18,7 +18,7 @@ spec:
|
||||
spec:
|
||||
containers:
|
||||
- name: llama-swap
|
||||
image: ghcr.io/mostlygeek/llama-swap:v198-vulkan-b8352
|
||||
image: ghcr.io/mostlygeek/llama-swap:v199-vulkan-b8576
|
||||
imagePullPolicy: IfNotPresent
|
||||
command:
|
||||
- /app/llama-swap
|
||||
|
||||
@@ -18,7 +18,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: open-webui
|
||||
version: 12.10.0
|
||||
version: 12.13.0
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: open-webui
|
||||
|
||||
@@ -15,7 +15,7 @@ spec:
|
||||
- name: renovate
|
||||
# Update this to the latest available and then enable Renovate on
|
||||
# the manifest
|
||||
image: renovate/renovate:43.76.1-full
|
||||
image: renovate/renovate:43.95.0-full
|
||||
envFrom:
|
||||
- secretRef:
|
||||
name: renovate-gitea-token
|
||||
|
||||
@@ -18,7 +18,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: cert-manager-webhook-ovh
|
||||
version: 0.9.4
|
||||
version: 0.9.5
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: cert-manager-webhook-ovh
|
||||
|
||||
@@ -23,7 +23,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: cert-manager
|
||||
version: v1.20.0
|
||||
version: v1.20.1
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: cert-manager
|
||||
|
||||
@@ -23,7 +23,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: cilium
|
||||
version: 1.19.1
|
||||
version: 1.19.2
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: cilium
|
||||
|
||||
@@ -23,7 +23,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: cloudnative-pg
|
||||
version: 0.27.1
|
||||
version: 0.28.0
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: cnpg
|
||||
|
||||
@@ -23,7 +23,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: k8up
|
||||
version: 4.8.6
|
||||
version: 4.9.0
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: k8up-io
|
||||
|
||||
@@ -23,7 +23,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: ingress-nginx
|
||||
version: 4.15.0
|
||||
version: 4.15.1
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: ingress-nginx
|
||||
|
||||
@@ -23,7 +23,7 @@ spec:
|
||||
chart:
|
||||
spec:
|
||||
chart: openbao
|
||||
version: 0.26.0
|
||||
version: 0.26.2
|
||||
sourceRef:
|
||||
kind: HelmRepository
|
||||
name: openbao
|
||||
|
||||
Reference in New Issue
Block a user