adithya-balaji commited on
Commit
1eff059
·
verified ·
1 Parent(s): 08d87bf

Add/update model card for medium

Browse files
Files changed (1) hide show
  1. README.md +7 -7
README.md CHANGED
@@ -21,12 +21,12 @@ This repo contains all four standard quantizations of this variant. **F16 is the
21
 
22
  | File | Quant | Size (MB) | Recall @ IoU 0.5 | Recall @ IoU 0.95 | Mean \|Δscore\| | Latency (median ms, T=8) |
23
  |---|---|---:|---:|---:|---:|---:|
24
- | `rfdetr-medium-f32.gguf` | F32 | 125.0 | 1.0000 | 1.0000 | 0.0008 | 139.9 |
25
- | `rfdetr-medium-f16.gguf` ← **recommended** | F16 | 67.2 | 1.0000 | 1.0000 | 0.0023 | 143.9 |
26
- | `rfdetr-medium-q8_0.gguf` | Q8_0 | 40.2 | 1.0000 | 1.0000 | 0.0063 | 145.8 |
27
- | `rfdetr-medium-q4_K.gguf` | Q4_K | 32.5 | 1.0000 | 1.0000 | 0.0230 | 155.8 |
28
 
29
- All accuracy numbers above are computed against the upstream PyTorch reference (`rfdetr 1.9.0`) on 1 image (test_image.jpg) at threshold 0.5. Latency is measured with `rfdetr-cli bench` (8 iters + 3 warmup) at T=8 threads on a single Intel Core i7-12800HX image (test_image.jpg).
30
 
31
  ## Architecture
32
 
@@ -71,7 +71,7 @@ build/bin/rfdetr-cli detect \
71
 
72
  ## Accuracy methodology
73
 
74
- All accuracy metrics are computed against the upstream PyTorch reference (`rfdetr 1.9.0`) on 1 image (test_image.jpg) at threshold 0.5. Each detection match uses greedy Hungarian-style assignment by IoU (≥ 0.5 lenient, ≥ 0.95 strict) with class equality required.
75
 
76
  See [BENCHMARK.md](https://github.com/adithyab94/rf-detr.cpp/blob/main/BENCHMARK.md) and [`benchmarks/results/accuracy_sweep.json`](https://github.com/adithyab94/rf-detr.cpp/blob/main/benchmarks/results/accuracy_sweep.json) for the full sweep across the (variant × quant) cells.
77
 
@@ -79,7 +79,7 @@ See [BENCHMARK.md](https://github.com/adithyab94/rf-detr.cpp/blob/main/BENCHMARK
79
 
80
  - Source project: [Roboflow RF-DETR](https://github.com/roboflow/rf-detr)
81
  - Upstream package: `rfdetr==1.9.0`
82
- - Converted with [rfdetr.cpp](https://github.com/adithyab94/rf-detr.cpp) at commit [`f0fd7f4ff643`](https://github.com/adithyab94/rf-detr.cpp/commit/f0fd7f4ff643824862071396cafeaf691f3cc800)
83
  - Checkpoint: official pretrained `rfdetr-medium` weights (downloaded by the `rfdetr` package on first use)
84
 
85
  ### Checksums (SHA-256)
 
21
 
22
  | File | Quant | Size (MB) | Recall @ IoU 0.5 | Recall @ IoU 0.95 | Mean \|Δscore\| | Latency (median ms, T=8) |
23
  |---|---|---:|---:|---:|---:|---:|
24
+ | `rfdetr-medium-f32.gguf` | F32 | 125.0 | 0.9731 | 0.9621 | 0.0024 | 139.9 |
25
+ | `rfdetr-medium-f16.gguf` ← **recommended** | F16 | 67.2 | 0.9780 | 0.9670 | 0.0022 | 143.9 |
26
+ | `rfdetr-medium-q8_0.gguf` | Q8_0 | 40.2 | 0.9890 | 0.9691 | 0.0070 | 145.8 |
27
+ | `rfdetr-medium-q4_K.gguf` | Q4_K | 32.5 | 0.8828 | 0.7390 | 0.0199 | 155.8 |
28
 
29
+ All accuracy numbers above are computed against the upstream PyTorch reference (`rfdetr 1.9.0`) on 7 images (000000000139.jpg, 000000000632.jpg, 000000039769.jpg, 000000087038.jpg, 000000252219.jpg, 000000397133.jpg, bus.jpg) at threshold 0.5. Latency is measured with `rfdetr-cli bench` (8 iters + 3 warmup) at T=8 threads on a single Intel Core i7-12800HX image (000000000139.jpg, 000000000632.jpg, 000000039769.jpg, 000000087038.jpg, 000000252219.jpg, 000000397133.jpg, bus.jpg).
30
 
31
  ## Architecture
32
 
 
71
 
72
  ## Accuracy methodology
73
 
74
+ All accuracy metrics are computed against the upstream PyTorch reference (`rfdetr 1.9.0`) on 7 images (000000000139.jpg, 000000000632.jpg, 000000039769.jpg, 000000087038.jpg, 000000252219.jpg, 000000397133.jpg, bus.jpg) at threshold 0.5. Each detection match uses greedy Hungarian-style assignment by IoU (≥ 0.5 lenient, ≥ 0.95 strict) with class equality required.
75
 
76
  See [BENCHMARK.md](https://github.com/adithyab94/rf-detr.cpp/blob/main/BENCHMARK.md) and [`benchmarks/results/accuracy_sweep.json`](https://github.com/adithyab94/rf-detr.cpp/blob/main/benchmarks/results/accuracy_sweep.json) for the full sweep across the (variant × quant) cells.
77
 
 
79
 
80
  - Source project: [Roboflow RF-DETR](https://github.com/roboflow/rf-detr)
81
  - Upstream package: `rfdetr==1.9.0`
82
+ - Converted with [rfdetr.cpp](https://github.com/adithyab94/rf-detr.cpp) at commit [`69d133e1e979`](https://github.com/adithyab94/rf-detr.cpp/commit/69d133e1e979c2dbf547507dfd8a5c5cec21e1f6)
83
  - Checkpoint: official pretrained `rfdetr-medium` weights (downloaded by the `rfdetr` package on first use)
84
 
85
  ### Checksums (SHA-256)