Commit Graph

Select branches

Hide Pull Requests

master

Mono Color

• f2ae80c9e5 Updated results with similarities master Rick McEwen 2026-03-31 12:30:14 -06:00
• 8b67b50d19 Add Burnley averaged embeddings test results to README Rick McEwen 2026-03-31 11:59:02 -06:00
• 5ce6265a90 Test data and results Rick McEwen 2026-03-31 11:54:39 -06:00
• 512f678310 Add latest test detection method Rick McEwen 2026-03-31 11:51:26 -06:00
• f598866d37 Add Burnley logo detection test using DetectLogosEmbeddings Rick McEwen 2026-03-31 11:49:11 -06:00
• 91d1c9cd59 Update README with recommended settings and test results Rick McEwen 2026-01-08 12:55:13 -05:00
• ea6fcec9ce Remove hybrid text+CLIP matching approach Rick McEwen 2026-01-08 12:48:39 -05:00
• f777b049a3 Fix EasyOCR model path to use script-relative directory Rick McEwen 2026-01-07 15:38:23 -05:00
• 49f982611a Add hybrid text+CLIP matching and image preprocessing Rick McEwen 2026-01-07 15:09:09 -05:00
• 78f46f04bf Add script to test optimal refs per logo for baseline CLIP Rick McEwen 2026-01-07 12:52:16 -05:00
• b5432c9ef7 Add comprehensive model comparison analysis Rick McEwen 2026-01-07 12:44:15 -05:00
• 440e8fcdb4 Combine all test results in a single directory Rick McEwen 2026-01-07 10:22:54 -05:00
• 2f28aa6052 Add threshold test script for image-split model Rick McEwen 2026-01-07 10:14:21 -05:00
• 569285f664 Use script directory as base path for portability Rick McEwen 2026-01-06 16:00:09 -05:00
• c086e8bbf7 Remove opencv-python from requirements (already installed) Rick McEwen 2026-01-06 15:23:31 -05:00
• 304d743df8 Add minimal requirements file for training server Rick McEwen 2026-01-06 15:17:34 -05:00
• 55abb1217c Add RTX 4090 config with image-level splits Rick McEwen 2026-01-06 14:23:13 -05:00
• 14a1bda3fa Add image-level split support for CLIP fine-tuning Rick McEwen 2026-01-05 15:10:45 -05:00
• 32bfefc022 Add threshold optimization script Rick McEwen 2026-01-05 14:20:27 -05:00
• f74d4b6981 Document threshold tuning for fine-tuned CLIP model Rick McEwen 2026-01-05 14:09:38 -05:00
• 6685af72d9 Add similarity distribution analysis for debugging embedding quality Rick McEwen 2026-01-05 13:39:20 -05:00
• 1bf9985def Fix double LoRA application when loading fine-tuned model Rick McEwen 2026-01-05 11:50:10 -05:00
• e5482a2d9e Add script to compare fine-tuned vs baseline CLIP Rick McEwen 2026-01-05 11:43:47 -05:00
• 99e5781c91 Fix trainer to use separation as sole criterion for best model Rick McEwen 2026-01-05 11:01:14 -05:00
• 44e8b6ae7d Add CLIP fine-tuning pipeline for logo recognition Rick McEwen 2026-01-04 13:45:25 -05:00
• 1551360028 Add embedding model comparison analysis (CLIP vs DINOv2) Rick McEwen 2026-01-02 16:26:59 -05:00
• 2c41549ae0 Document margin behavior and update model comparison script Rick McEwen 2026-01-02 14:42:53 -05:00
• 48d9145810 Update README with model selection and new test scripts Rick McEwen 2026-01-02 12:53:50 -05:00
• 2d19ed91d7 Document mean vs max similarity aggregation in multi-ref matching Rick McEwen 2026-01-02 12:17:13 -05:00
• 94db5bd40b Add embedding model selection and comparison test scripts Rick McEwen 2026-01-02 12:05:27 -05:00
• a3008ee57f Remove extraneous file from repository, keep local only Rick McEwen 2025-12-31 17:53:06 -05:00
• ea589a50a4 Update README with new test parameters and dataset setup Rick McEwen 2025-12-31 17:49:56 -05:00
• 41c75356d9 Add --output-file option for clean results output Rick McEwen 2025-12-31 17:42:52 -05:00
• 41bc0c701f Add simple matching method as baseline for comparison tests Rick McEwen 2025-12-31 17:36:18 -05:00
• 197e007591 Add margin check to multi-ref matching to reduce false positives Rick McEwen 2025-12-31 11:23:47 -05:00
• ddccf653d2 Initial commit: Logo detection test framework Rick McEwen 2025-12-31 10:42:36 -05:00