minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[docs] Inkling cookbook: LoRA cells require --disable-prefill-cuda-graph (#31418)

Browse Source
This commit is contained in:
Yanbin Jiang
2026-07-16 05:19:49 +00:00
committed by GitHub
parent 0b372d03de
commit 40517b593b
1 changed files with 7 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs_new/src/snippets/configs/thinkingmachines/inkling.jsx
+7
View File
@@ -871,6 +871,7 @@ export const config = {
"--reasoning-parser inkling",
"--tool-call-parser inkling",
"--enable-lora",
"--disable-prefill-cuda-graph",
"--lora-backend triton",
"--lora-use-virtual-experts",
"--max-loras-per-batch {{MAX_LORAS}}",
@@ -904,6 +905,7 @@ export const config = {
"--reasoning-parser inkling",
"--tool-call-parser inkling",
"--enable-lora",
"--disable-prefill-cuda-graph",
"--lora-backend triton",
"--lora-use-virtual-experts",
"--max-loras-per-batch {{MAX_LORAS}}",
@@ -937,6 +939,7 @@ export const config = {
"--reasoning-parser inkling",
"--tool-call-parser inkling",
"--enable-lora",
"--disable-prefill-cuda-graph",
"--lora-backend triton",
"--lora-use-virtual-experts",
"--max-loras-per-batch {{MAX_LORAS}}",
@@ -970,6 +973,7 @@ export const config = {
"--reasoning-parser inkling",
"--tool-call-parser inkling",
"--enable-lora",
"--disable-prefill-cuda-graph",
"--lora-backend triton",
"--lora-use-virtual-experts",
"--max-loras-per-batch {{MAX_LORAS}}",
@@ -1004,6 +1008,7 @@ export const config = {
"--reasoning-parser inkling",
"--tool-call-parser inkling",
"--enable-lora",
"--disable-prefill-cuda-graph",
"--lora-backend triton",
"--lora-use-virtual-experts",
"--max-loras-per-batch {{MAX_LORAS}}",
@@ -1040,6 +1045,7 @@ export const config = {
"--reasoning-parser inkling",
"--tool-call-parser inkling",
"--enable-lora",
"--disable-prefill-cuda-graph",
"--lora-backend triton",
"--lora-use-virtual-experts",
"--max-loras-per-batch {{MAX_LORAS}}",
@@ -1071,6 +1077,7 @@ export const config = {
"--reasoning-parser inkling",
"--tool-call-parser inkling",
"--enable-lora",
"--disable-prefill-cuda-graph",
"--lora-backend triton",
"--lora-use-virtual-experts",
"--max-loras-per-batch {{MAX_LORAS}}",