Skip to content

GitLab

Unverified Commit f8208f8d authored by Graham King's avatar Graham King Committed by GitHub
Browse files

chore: Pin YAML pip installs to reduce supply chain risk (#8306) 


Signed-off-by: default avatarGraham King <grahamk@nvidia.com>
parent 876aeb35
Hide whitespace changes
Inline Side-by-side
Showing with 42 additions and 43 deletions
examples/backends/vllm/deploy/lora/multimodal/sync-lora-job.yaml
View file @ f8208f8d
......@@ -15,7 +15,7 @@ spec:
- -c
- |
set -eux
pip install --no-cache-dir huggingface-hub awscli
pip install --no-cache-dir huggingface-hub==1.11.0 awscli==1.44.80
hf download $MODEL_NAME --local-dir /tmp/lora
rm -rf /tmp/lora/.cache
aws --endpoint-url=http://minio:9000 s3 mb s3://$LORA_ROOT_PATH || true
......@@ -35,4 +35,4 @@ spec:
- name: MODEL_NAME
value: Chhagan005/Chhagan-DocVL-Qwen3
restartPolicy: Never
backoffLimit: 3
\ No newline at end of file
backoffLimit: 3
examples/backends/vllm/deploy/lora/sync-lora-job.yaml
View file @ f8208f8d
......@@ -15,7 +15,7 @@ spec:
- -c
- |
set -eux
pip install --no-cache-dir huggingface-hub awscli
pip install --no-cache-dir huggingface-hub==1.11.0 awscli==1.44.80
hf download $MODEL_NAME --local-dir /tmp/lora
rm -rf /tmp/lora/.cache
aws --endpoint-url=http://minio:9000 s3 mb s3://$LORA_ROOT_PATH || true
......@@ -35,4 +35,4 @@ spec:
- name: MODEL_NAME
value: codelion/Qwen3-0.6B-accuracy-recovery-lora
restartPolicy: Never
backoffLimit: 3
\ No newline at end of file
backoffLimit: 3
recipes/deepseek-r1/model-cache/model-download-sglang.yaml
View file @ f8208f8d
......@@ -20,14 +20,14 @@ spec:
image: python:3.10-slim
command: ["sh", "-c"]
env:
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
- name: HF_HOME
value: /opt/model-cache
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download deepseek-ai/DeepSeek-R1
volumeMounts:
- name: model-cache
......@@ -35,4 +35,4 @@ spec:
volumes:
- name: model-cache
persistentVolumeClaim:
claimName: model-cache
\ No newline at end of file
claimName: model-cache
recipes/deepseek-r1/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -20,7 +20,7 @@ spec:
image: python:3.10-slim
command: ["sh", "-c"]
env:
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
# Optional: create with: kubectl create secret generic hf-token-secret --from-literal=HF_TOKEN=<token> -n <namespace>
- name: HF_TOKEN
......@@ -32,7 +32,7 @@ spec:
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download nvidia/DeepSeek-R1-FP4 --local-dir /model-cache/deepseek-r1-fp4
hf download deepseek-ai/DeepSeek-R1 --local-dir /model-cache/deepseek-r1
volumeMounts:
......@@ -41,4 +41,4 @@ spec:
volumes:
- name: model-cache
persistentVolumeClaim:
claimName: model-cache
\ No newline at end of file
claimName: model-cache
recipes/deepseek-r1/trtllm/disagg/wide_ep/gb200/perf.yaml
View file @ f8208f8d
......@@ -30,7 +30,7 @@ spec:
- -c
- |
apt-get update && apt-get install -y curl jq procps git && apt-get clean
pip install aiperf;
pip install aiperf==0.6.0;
echo "aiperf installation completed";
sysctl -w net.ipv4.ip_local_port_range="1024 65000"
cat /proc/sys/net/ipv4/ip_local_port_range
......@@ -152,4 +152,3 @@ spec:
- name: model-cache
persistentVolumeClaim:
claimName: model-cache
recipes/deepseek-v32-fp4/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -32,12 +32,12 @@ spec:
value: nvidia/DeepSeek-V3.2-NVFP4
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME
volumeMounts:
- name: model-cache
......
recipes/glm-5-nvfp4/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -31,12 +31,12 @@ spec:
value: nvidia/GLM-5-NVFP4
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME
volumeMounts:
- name: model-cache
......
recipes/glm-5-nvfp4/sglang/disagg/perf.yaml
View file @ f8208f8d
......@@ -53,7 +53,7 @@ spec:
- |
set -eu
apt-get update -qq && apt-get install -y -qq curl jq && apt-get clean
pip install -q aiperf transformers tokenizers
pip install -q aiperf==0.6.0 transformers==4.57.3 tokenizers==0.22.2
echo "Waiting for model at http://$ENDPOINT/v1/models..."
while ! curl -sf "http://$ENDPOINT/v1/models" | jq -e --arg m "$TARGET_MODEL" '.data[]? | select(.id == $m)' >/dev/null 2>&1; do
......
recipes/gpt-oss-120b/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -26,14 +26,14 @@ spec:
value: openai/gpt-oss-120b
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
- name: MODEL_REVISION
value: b5c939de8f754692c1647ca79fbf85e8c1e70f8a
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME --revision $MODEL_REVISION --exclude "original/*" --exclude "metal/*"
volumeMounts:
- name: model-cache
......@@ -41,4 +41,4 @@ spec:
volumes:
- name: model-cache
persistentVolumeClaim:
claimName: model-cache
\ No newline at end of file
claimName: model-cache
recipes/kimi-k2.5/model-cache/baseten/model-download.yaml
View file @ f8208f8d
......@@ -26,12 +26,12 @@ spec:
value: baseten-admin/Kimi-2.5-text-nvfp4-v3
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME
volumeMounts:
- name: model-cache
......@@ -39,4 +39,4 @@ spec:
volumes:
- name: model-cache
persistentVolumeClaim:
claimName: model-cache
\ No newline at end of file
claimName: model-cache
recipes/kimi-k2.5/model-cache/nvidia/eagle-download.yaml
View file @ f8208f8d
......@@ -28,12 +28,12 @@ spec:
value: 0b0c6ac039089ad2c2418c91c039553381a302d9
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download "$MODEL_NAME" --revision "$MODEL_REVISION"
volumeMounts:
- name: model-cache
......
recipes/kimi-k2.5/model-cache/nvidia/model-download.yaml
View file @ f8208f8d
......@@ -26,12 +26,12 @@ spec:
value: nvidia/Kimi-K2.5-NVFP4
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME
volumeMounts:
- name: model-cache
......@@ -39,4 +39,4 @@ spec:
volumes:
- name: model-cache
persistentVolumeClaim:
claimName: model-cache
\ No newline at end of file
claimName: model-cache
recipes/llama-3-70b/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -26,14 +26,14 @@ spec:
value: "RedHatAI/Llama-3.3-70B-Instruct-FP8-dynamic"
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
- name: MODEL_REVISION
value: ddb4128556dfcff99e0c41aee159ea6c3e655dcd
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME --revision $MODEL_REVISION
volumeMounts:
- name: model-cache
......@@ -41,4 +41,4 @@ spec:
volumes:
- name: model-cache
persistentVolumeClaim:
claimName: model-cache
\ No newline at end of file
claimName: model-cache
recipes/nemotron-3-super-fp8/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -26,12 +26,12 @@ spec:
value: nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME
volumeMounts:
- name: model-cache
......
recipes/qwen3-235b-a22b-fp8/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -26,14 +26,14 @@ spec:
value: Qwen/Qwen3-235B-A22B-FP8
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
- name: MODEL_REVISION
value: 39eb2b067ea6b8e3e1dd97d3cd0c7ffeaf3e1a35
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME --revision $MODEL_REVISION
volumeMounts:
- name: model-cache
......
recipes/qwen3-32b-fp8/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -26,14 +26,14 @@ spec:
value: Qwen/Qwen3-32B-FP8
- name: HF_HOME
value: /model-store
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
- name: MODEL_REVISION
value: aa55da1ecc13d006e8b8e4f54579b1ea8c3db2df
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME --revision $MODEL_REVISION
volumeMounts:
- name: model-cache
......
recipes/qwen3-32b/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -26,14 +26,14 @@ spec:
value: "Qwen/Qwen3-32B"
- name: HF_HOME
value: /home/dynamo/.cache/huggingface
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
- name: MODEL_REVISION
value: 9216db5781bf21249d130ec9da846c4624c16137
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download $MODEL_NAME --revision $MODEL_REVISION
volumeMounts:
- name: model-cache
......
recipes/qwen3-32b/vllm/agg-round-robin/perf.yaml
View file @ f8208f8d
......@@ -20,7 +20,7 @@ spec:
apt update && apt install tmux wget curl jq -y
# Install benchmarking tool
pip install aiperf
pip install aiperf==0.6.0
# Wait for model to be ready
echo "Waiting for model '${MODEL_NAME}' at http://${FRONTEND}:8000/v1/models..."
......
recipes/qwen3-32b/vllm/disagg-kv-router/perf.yaml
View file @ f8208f8d
......@@ -20,7 +20,7 @@ spec:
apt update && apt install tmux wget curl jq -y
# Install benchmarking tool
pip install aiperf
pip install aiperf==0.6.0
# Wait for model to be ready
echo "Waiting for model '${MODEL_NAME}' at http://${FRONTEND}:8000/v1/models..."
......
recipes/qwen3-vl-30b/model-cache/model-download.yaml
View file @ f8208f8d
......@@ -26,14 +26,14 @@ spec:
value: "Qwen/Qwen3-VL-30B-A3B-Instruct-FP8" # Remove FP8 for BF16 variant
- name: HF_HOME
value: /home/dynamo/.cache/huggingface
- name: HF_HUB_ENABLE_HF_TRANSFER
- name: HF_XET_HIGH_PERFORMANCE
value: "1"
- name: MODEL_REVISION
value: "main"
args:
- |
set -eux
pip install --no-cache-dir huggingface_hub hf_transfer
pip install --no-cache-dir huggingface_hub==1.11.0
hf download "$MODEL_NAME" --revision "$MODEL_REVISION"
volumeMounts:
- name: model-cache
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment