remove add_bias option

[Limerances]

Motivation

Remove the add_bias option; this option is no longer supported.

Modifications

Usage or Command

Accuracy Tests

Checklist

• Add at least a tag in the PR title.
  • Tag list: [[FDConfig],[APIServer],[Engine], [Scheduler], [PD Disaggregation], [Executor], [Graph Optimization], [Speculative Decoding], [RL], [Models], [Quantization], [Loader], [OP], [KVCache], [DataProcessor], [BugFix], [Docs], [CI], [Optimization], [Feature], [Benchmark], [Others], [XPU], [HPU], [GCU], [DCU], [Iluvatar], [Metax]]
  • You can add new tags based on the PR content, but the semantics must be clear.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
• Provide accuracy results.
• If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

[paddle-bot]

Thanks for your contribution!

[Copilot]

Pull request overview

This PR removes the deprecated add_bias option from linear layers throughout the codebase. The parameter was previously used to control whether bias should be added in the current layer or in pre/post layers, but is no longer supported.

Key changes:

• Removed add_bias parameter from all linear layer classes (LinearBase, ColumnParallelLinear, MergedColumnParallelLinear, ReplicatedLinear, GatedLinear, QKVParallelLinear, RowParallelLinear)
• Updated quantization methods to unconditionally pass layer.bias instead of checking add_bias flag
• Cleaned up test mocks that referenced the removed parameter

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
fastdeploy/model_executor/layers/linear.py	Removed add_bias parameter from all linear layer class constructors and docstrings; simplified bias handling in RowParallelLinear
fastdeploy/model_executor/layers/quantization/weight_only.py	Changed bias parameter to always pass layer.bias instead of conditional based on add_bias
fastdeploy/model_executor/layers/quantization/w4afp8.py	Changed bias parameter to always pass layer.bias instead of conditional based on add_bias
fastdeploy/model_executor/models/gpt_oss.py	Removed add_bias=True argument from RowParallelLinear instantiation
tests/quantization/test_w4afp8.py	Removed mock setup of add_bias attribute from test fixtures

Comments suppressed due to low confidence (1)

tests/quantization/test_w4afp8.py:146

• After removing the add_bias parameter, this test no longer properly tests the "without_bias" scenario. The setUp method sets self.layer.bias = "bias" (line 43), so this test now actually tests with a bias present.

To properly test the case when bias is None, you should add:

self.layer.bias = None

at the beginning of this test method (after line 141), so that it correctly verifies the behavior when no bias is provided.

    def test_apply_without_bias(self, mock_gemm):
        mock_gemm.return_value = "out"
        x = "x"

        result = self.method.apply(self.layer, x)
        self.assertEqual(result, "out")
        args = mock_gemm.call_args.kwargs
        self.assertIsNone(args["bias"])

[Copilot]

The code now unconditionally accesses self.bias when self.tp_size > 1 and self.reduce_results, but self.bias is None when with_bias=False (see line 173-179 in LinearLayer.init). This will cause an AttributeError when trying to set weight attributes on a None object.

Consider adding a condition to check if bias exists:

if self.tp_size > 1 and self.reduce_results and self.with_bias:
    set_weight_attrs(self.bias, {"tp_row_bias": True})

Suggested change

	if self.tp_size > 1 and self.reduce_results:
	if self.tp_size > 1 and self.reduce_results and with_bias:

[codecov-commenter]

Codecov Report

❌ Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@3066a0c). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/model_executor/layers/linear.py	0.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #5425   +/-   ##
==========================================
  Coverage           ?   59.51%           
==========================================
  Files              ?      327           
  Lines              ?    40643           
  Branches           ?     6170           
==========================================
  Hits               ?    24188           
  Misses             ?    14588           
  Partials           ?     1867           

Flag	Coverage Δ
GPU	59.51% <0.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.