Utsab Dahal

Fix TypeError in fin_quant coding step (#1224)

3

This PR fixes the TypeError: data must be str, not dict issue reported in #1224. I modified the `assign_code_list_to_evo` function in `factor_coder/evolving_strategy.py` and `model_coder/evolving_strategy.py` to handle dictionary responses from the...

## GPU Support for Time Series Models **Resolves #1256** ### Problem Time series models (LSTM) in Co-STEER framework were CPU-only, causing significant resource waste. ### Solution Implement automatic GPU detection...

Fix: handle 3D input for channels_first in preprocess_input() #21749

3

### Summary Fixed a shape mismatch issue when `preprocess_input()` receives 3D input (single image) with `data_format='channels_first'`. The bug occurred because the mean and std tensors were reshaped for 4D input...

Fix: Remove deprecated .path access in Muon optimizer for TF 2.16+ compatibility

5

### Summary Fixes keras.optimizers.Muon failing with AttributeError: 'ResourceVariable' object has no attribute 'path' in Keras 3 / TF 2.16–2.20. ### Changes - Replaced deprecated `.path` references with `.name` for variable...

Fix: correct effective crop_to_aspect_ratio logic and lint issues

7

## PR Description This PR fixes the logic controlling crop_to_aspect_ratio inside transform_images() in the Resizing layer. ### Summary of changes: Ensures cropping only occurs when aspect ratios differ. Adds a...

Fix/random_crop validation behavior

3

# Fix RandomCrop validation behavior and edge cases Fixes #21868 ## Problem Description The `RandomCrop` layer had two critical issues: 1. **During validation/inference, the layer did nothing** - Images were...