Compare commits
1 Commits
94ea73eb61
...
7a20ecd526
| Author | SHA1 | Date | |
|---|---|---|---|
| 7a20ecd526 |
@@ -6,7 +6,6 @@ macros:
|
||||
base_args: "--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 q8_0 -ctv q8_0"
|
||||
@@ -164,7 +163,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-4B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_128k}
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -173,7 +172,7 @@ models:
|
||||
cmd: |
|
||||
/app/llama-server
|
||||
-hf unsloth/Qwen3.5-4B-GGUF:Q4_K_M
|
||||
${qwen35_ctx_128k}
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_off}
|
||||
@@ -183,7 +182,7 @@ models:
|
||||
/app/llama-server
|
||||
-hf mradermacher/Qwen3.5-4B-heretic-GGUF:Q4_K_M
|
||||
${qwen35_4b_heretic_mmproj}
|
||||
${qwen35_ctx_128k}
|
||||
${qwen35_ctx_256k}
|
||||
${qwen35_sampling}
|
||||
${common_args}
|
||||
${thinking_on}
|
||||
@@ -193,7 +192,7 @@ models:
|
||||
/app/llama-server
|
||||
-hf mradermacher/Qwen3.5-4B-heretic-GGUF:Q4_K_M
|
||||
${qwen35_4b_heretic_mmproj}
|
||||
${qwen35_ctx_128k}
|
||||
${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:v199-vulkan-b8576
|
||||
image: ghcr.io/mostlygeek/llama-swap:v199-vulkan-b8562
|
||||
imagePullPolicy: IfNotPresent
|
||||
command:
|
||||
- /app/llama-swap
|
||||
|
||||
@@ -15,7 +15,7 @@ spec:
|
||||
- name: renovate
|
||||
# Update this to the latest available and then enable Renovate on
|
||||
# the manifest
|
||||
image: renovate/renovate:43.101.2-full
|
||||
image: renovate/renovate:43.99.1-full
|
||||
envFrom:
|
||||
- secretRef:
|
||||
name: renovate-gitea-token
|
||||
|
||||
Reference in New Issue
Block a user