jens/pyRBM
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[README] - update faces_sub_image with training results and CLI args

Browse Source 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
jens
2026-05-30 21:15:02 +02:00
co-authored by Claude Sonnet 4.6
parent 68b376657c
commit 8c10bda002
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/tests/README.md
+1 -1
View File
@@ -29,7 +29,7 @@ These load images from a fixed path on disk.
| File | RBM type | What it tests | | File | RBM type | What it tests |
|---|---|---| |---|---|---|
| `test_faces_sub_image.py` | GB-RBM | Auto-encoder on 32×32 patches from Caltech WebFaces (`/media/jens/cifs/bilder/MachineVision/Caltech_WebFaces/`). Uses `SubImage` with 50% overlap (stride=16); reconstruction blends overlapping patches by pixel-averaging. Supports RGB and grayscale modes. CLI args: `--grayscale`, `--load_model`, `--do_train`, `--l1_lambda`. After training: shows learned weight filters, patch-level reconstructions, and a full overlap-blended image reconstruction. Default: 256 hidden units, 1000 epochs, mini_batch=1000. **L1=0.01 is the recommended sweet spot** — produces structured edge/texture filters with 15% sparsity while keeping MSE~0.242. L1=0.05 collapses most filters to blob detectors (58% sparsity, MSE~0.261). | | `test_faces_sub_image.py` | GB-RBM | Auto-encoder on 32×32 patches from Caltech WebFaces (`/media/jens/cifs/bilder/MachineVision/Caltech_WebFaces/`). Uses `SubImage` with 50% overlap (stride=16); reconstruction blends overlapping patches by pixel-averaging. Supports RGB and grayscale modes. CLI args: `--grayscale`, `--load_model`, `--do_train`, `--l1_lambda`, `--num_epochs`, `--mini_batch_size`. After training: shows learned weight filters, patch-level reconstructions, and a full overlap-blended image reconstruction. Results summary (grayscale, 256 hidden units, 1000 epochs): L1=0.01 → MSE 0.168, MAE 0.214, structured edge/texture filters; L1=0.05 → MSE 0.195, MAE 0.228, sparse blob detectors. **L1=0.01 is the recommended sweet spot.** Mini-batch size has little effect on final quality (mini_batch=200 vs 1000 give equivalent MSE) but smaller batches amplify the effective L1 penalty per epoch due to more frequent updates. |
| `test_sub_image.py` | — | Unit test for `SubImage` (patch extraction). No RBM involved. | | `test_sub_image.py` | — | Unit test for `SubImage` (patch extraction). No RBM involved. |
| `test_conv2d.py` | — | Unit test for `conv2d`. No RBM involved. | | `test_conv2d.py` | — | Unit test for `conv2d`. No RBM involved. |