Skip to content

v0.3.2 #19

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
RogerLiu312 merged 1 commit into from
Nov 10, 2026
Merged

v0.3.2 #19

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions examples/self_contained/docker/sglang_glm52.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,10 +29,10 @@ version: "0.1"
variables:
HOST_ADDRESS:
description: "Docker host that runs both tasks (ssh-reachable from the driver)"
value: 10.114.98.206
value: 10.0.0.11
SSH_USER:
description: "SSH login on that host; leave empty to use your own"
value: rogliu
value: ""

SERVED_MODEL_NAME:
description: "Name the server advertises; aiperf must request the same string"
Expand Down Expand Up @@ -107,7 +107,7 @@ variables:
(/root/.cache/sglang/flashinfer/autotune/<ver>/<sm>/...) survive the --rm
container and later runs skip the rebuild. Entries are keyed by version and
arch, so one dir can serve several images. Docker creates it if missing.
value: /var/data0/sandbox/rogliu/flashinfer-cache
value: /var/tmp/sflow-sglang-cache

SGLANG_IMAGE:
description: "SGLang image new enough for GLM-5.2 and these flags; pin a tag for reproducible runs"
Expand All @@ -118,7 +118,7 @@ artifacts:
# the server loads weights from it, aiperf loads the tokenizer from it.
- name: MODEL_PATH
description: "GLM-5.2 NVFP4 checkpoint, a path on the docker host"
uri: fs:///var/data0/Games/Huggingface_Models_nvidia_GLM_5_2_NVFP4_1
uri: fs:///path/to/GLM-5.2-NVFP4

backends:
- name: docker_host
Expand Down
2 changes: 1 addition & 1 deletion examples/self_contained/docker/vllm_disagg_glm53.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,7 +186,7 @@ variables:
artifacts:
- name: LOCAL_MODEL_PATH
description: "GLM-5.3-Flash NVFP4 weights, present on every host at this path"
uri: fs:///var/data0/sandbox/rogliu/nvidia-GLM-5.3-Flash-NVFP4
uri: fs:///path/to/GLM-5.3-Flash-NVFP4

backends:
- name: docker_cluster
Expand Down
8 changes: 4 additions & 4 deletions src/sflow/samples/self_contained/docker/sglang_glm52.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,10 +29,10 @@ version: "0.1"
variables:
HOST_ADDRESS:
description: "Docker host that runs both tasks (ssh-reachable from the driver)"
value: 10.114.98.206
value: 10.0.0.11
SSH_USER:
description: "SSH login on that host; leave empty to use your own"
value: rogliu
value: ""

SERVED_MODEL_NAME:
description: "Name the server advertises; aiperf must request the same string"
Expand Down Expand Up @@ -107,7 +107,7 @@ variables:
(/root/.cache/sglang/flashinfer/autotune/<ver>/<sm>/...) survive the --rm
container and later runs skip the rebuild. Entries are keyed by version and
arch, so one dir can serve several images. Docker creates it if missing.
value: /var/data0/sandbox/rogliu/flashinfer-cache
value: /var/tmp/sflow-sglang-cache

SGLANG_IMAGE:
description: "SGLang image new enough for GLM-5.2 and these flags; pin a tag for reproducible runs"
Expand All @@ -118,7 +118,7 @@ artifacts:
# the server loads weights from it, aiperf loads the tokenizer from it.
- name: MODEL_PATH
description: "GLM-5.2 NVFP4 checkpoint, a path on the docker host"
uri: fs:///var/data0/Games/Huggingface_Models_nvidia_GLM_5_2_NVFP4_1
uri: fs:///path/to/GLM-5.2-NVFP4

backends:
- name: docker_host
Expand Down
2 changes: 1 addition & 1 deletion src/sflow/samples/self_contained/docker/vllm_disagg_glm53.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,7 +186,7 @@ variables:
artifacts:
- name: LOCAL_MODEL_PATH
description: "GLM-5.3-Flash NVFP4 weights, present on every host at this path"
uri: fs:///var/data0/sandbox/rogliu/nvidia-GLM-5.3-Flash-NVFP4
uri: fs:///path/to/GLM-5.3-Flash-NVFP4

backends:
- name: docker_cluster
Expand Down
Loading