[CI] Shard Quantization job into 4 parallel shards (≤30 min target) (#52328)
This commit is contained in:
@@ -2,10 +2,11 @@ group: Quantization
|
|||||||
depends_on:
|
depends_on:
|
||||||
- image-build
|
- image-build
|
||||||
steps:
|
steps:
|
||||||
- label: Quantization
|
- label: Quantization %N
|
||||||
device: h200_35gb
|
device: h200_35gb
|
||||||
key: quantization
|
key: quantization
|
||||||
timeout_in_minutes: 75
|
timeout_in_minutes: 40
|
||||||
|
parallelism: 4
|
||||||
env:
|
env:
|
||||||
VLLM_USE_V2_MODEL_RUNNER: "0"
|
VLLM_USE_V2_MODEL_RUNNER: "0"
|
||||||
source_file_dependencies:
|
source_file_dependencies:
|
||||||
@@ -18,7 +19,7 @@ steps:
|
|||||||
- uv pip install --system conch-triton-kernels
|
- uv pip install --system conch-triton-kernels
|
||||||
# The SM90-only checkpoint currently contains a removed weight_chan_scale
|
# The SM90-only checkpoint currently contains a removed weight_chan_scale
|
||||||
# parameter. It was not exercised by the previous L4 job.
|
# parameter. It was not exercised by the previous L4 job.
|
||||||
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py -k 'not test_compressed_tensors_w4a8_fp8'
|
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py -k 'not test_compressed_tensors_w4a8_fp8' --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||||
|
|
||||||
- label: Quantized Fusions
|
- label: Quantized Fusions
|
||||||
device: h200_35gb
|
device: h200_35gb
|
||||||
|
|||||||
Reference in New Issue
Block a user