Fix: Respect inference_mode when setting adapters with modules_to_save (Issue #2928)

[ada-ggf25]

Fix: Respect inference_mode when setting adapters with modules_to_save

Fixes #2928

Description

This PR fixes issue #2928 where modules_to_save had requires_grad=True even when inference_mode=True was passed to set_adapter() caused issues when using quantized models (e.g., with bitsandbytes) in inference mode, as the quantized layers require parameters to have requires_grad=False.

Problem

When calling model.set_adapter(adapter_name, inference_mode=True) with a model that has modules_to_save configured (e.g., a classifichead), the modules_to_save parameters would still have requires_grad=True despite being in inference mode. This happened because:

1. _set_adapter() was calling module.enable_adapters(True) unconditionally
2. ModulesToSaveWrapper.enable_adapters(True) sets requires_grad_(True) for all active adapters
3. This occurred before set_adapter() was called with inference_mode, creating a conflict

Solution

The fix ensures that enable_adapters() is only called when not in inference mode:

1. Modified _set_adapter() in src/peft/utils/other.py:

   • Added conditional check: only call enable_adapters(True) when inference_mode=False
   • This prevents setting gradients to True when inference mode should keep them False

2. Updated PeftModel.set_adapter() in src/peft/peft_model.py:

   • Added inference_mode parameter to match the API of PeftMixedModel.set_adapter()
   • Passes inference_mode to both _set_adapter() and base_model.set_adapter()

3. Added comprehensive tests in tests/test_other.py:

   • test_modules_to_save_inference_mode_requires_grad_false: Verifies requires_grad=False in inference mode
   • test_modules_to_save_training_mode_requires_grad_true: Verifies requires_grad=True in training mode
   • test_modules_to_save_inference_mode_with_torch_inference_mode: Verifies compatibility with torch.inference_mode()

Changes Made

Code Changes

src/peft/utils/other.py:

• Modified _set_adapter() to conditionally call enable_adapters() based on inference_mode parameter

src/peft/peft_model.py:

• Updated set_adapter() method signature to accept inference_mode parameter
• Passes inference_mode to underlying adapter setting functions

tests/test_other.py:

• Added TestModulesToSaveInferenceMode test class with 3 comprehensive tests

Testing

Test Results

All new tests pass:

• test_modules_to_save_inference_mode_requires_grad_false - PASSED
• test_modules_to_save_training_mode_requires_grad_true - PASSED
• test_modules_to_save_inference_mode_with_torch_inference_mode - PASSED

All existing modules_to_save tests pass (11/11)

Related tests pass (71/74 - 3 failures are unrelated BOFT dependency issues)

Code quality checks pass (make quality)

Test Coverage

The new tests verify:

• modules_to_save parameters correctly have requires_grad=False when inference_mode=True
• modules_to_save parameters correctly have requires_grad=True when inference_mode=False (training mode)
• Compatibility with torch.inference_mode() context manager

Example Usage

Before this fix, the following code would fail with quantized models:

model = PeftModel.from_pretrained(base_model, adapter_path)
model = convert_to_int8(model)  # Quantization
model.eval()

with torch.inference_mode():
    model.set_adapter("my_adapter", inference_mode=True)  #  modules_to_save still had requires_grad=True
    _ = model(batch)

After this fix:

model = PeftModel.from_pretrained(base_model, adapter_path)
model = convert_to_int8(model)  # Quantization
model.eval()

with torch.inference_mode():
    model.set_adapter("my_adapter", inference_mode=True)  #  modules_to_save correctly have requires_grad=False
    _ = model(batch)