Skip to content

Fix some CI issues#1842

Open
lvliang-intel wants to merge 9 commits into
mainfrom
lvl/fix_ci_issues
Open

Fix some CI issues#1842
lvliang-intel wants to merge 9 commits into
mainfrom
lvl/fix_ci_issues

Conversation

@lvliang-intel
Copy link
Copy Markdown
Contributor

Description

Fix some CI issues:
https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_apis/build/builds/63619/logs/68
https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_apis/build/builds/63619/logs/62
https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_apis/build/builds/63619/logs/76

image

Type of Change

Bug fix

Related Issues

Fixes or relates to #

Checklist Before Submitting

  • My code has been tested locally.
  • Documentation has been updated as needed.
  • New or updated tests are included where applicable.
  • The CUDA CI has passed. You can trigger it by commenting /azp run Unit-Test-CUDA-AutoRound.

Signed-off-by: lvliang-intel <liang1.lv@intel.com>
Copilot AI review requested due to automatic review settings May 21, 2026 08:12
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

This PR addresses CI failures by making model path resolution in tests more robust, hardening GGUF export conversion against missing model attributes, and refining GGUF-format detection / ignore-layer handling in the compressor layer config logic.

Changes:

  • Update GGUF export test to resolve the model via a helper path resolver.
  • Guard GGUF tensor conversion against missing _checkpoint_conversion_mapping.
  • Refine GGUF format detection and adjust ignore-layer preprocessing in configure_layer_config.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File Description
test/test_cpu/export/test_gguf_format.py Uses a model-path helper for more reliable CI execution.
auto_round/export/export_to_gguf/convert.py Avoids attribute errors when conversion mapping is absent.
auto_round/compressors/base.py Updates GGUF format detection and ignore-layer preprocessing logic.

cls.model.tensor_name_list = []

reverse_key_mapping = {v: k for k, v in cls.model._checkpoint_conversion_mapping.items()}
reverse_key_mapping = {v: k for k, v in getattr(cls.model, "_checkpoint_conversion_mapping", {}).items()}
is_gguf_format = (f := getattr(self.compress_context, "formats", None)) is not None and "gguf" in f
_formats = getattr(self.compress_context, "formats", None)
is_gguf_format = _formats is not None and any(
"gguf" in str(getattr(fmt, "output_format", "")) for fmt in _formats
Comment on lines 563 to +577
predefined_ignore_layers = get_predefined_ignore_layers(self.model_context.model)
compressed_predefined_ignore_layers = compress_layer_names(predefined_ignore_layers)
if not is_gguf_format:
predefined_ignore_layers = get_predefined_ignore_layers(self.model_context.model)
if predefined_ignore_layers and self.quant_block_list:
block_prefixes = [block for group in self.quant_block_list for block in group]
predefined_ignore_layers = [
name
for name in predefined_ignore_layers
if any(name.startswith(prefix) for prefix in block_prefixes)
]
predefined_ignore_layers = compress_layer_names(predefined_ignore_layers)
if predefined_ignore_layers:
logger.info(f"Using predefined ignore_layers: {compressed_predefined_ignore_layers}")
tmp_str = ",".join(predefined_ignore_layers)
tmp_str = predefined_ignore_layers.replace(" ", "")
lvliang-intel and others added 5 commits May 21, 2026 21:47
Signed-off-by: lvliang-intel <liang1.lv@intel.com>
… into lvl/fix_ci_issues

Signed-off-by: lvliang-intel <liang1.lv@intel.com>
@lvliang-intel
Copy link
Copy Markdown
Contributor Author

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

Signed-off-by: lvliang-intel <liang1.lv@intel.com>
@lvliang-intel
Copy link
Copy Markdown
Contributor Author

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@lvliang-intel
Copy link
Copy Markdown
Contributor Author

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@chensuyue chensuyue added this to the 0.13.0 milestone May 22, 2026
Signed-off-by: lvliang-intel <liang1.lv@intel.com>
@lvliang-intel
Copy link
Copy Markdown
Contributor Author

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants