Instructions to use thangkt/PCB-Prune-YOLO-P40-A8-Direct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- ultralytics
How to use thangkt/PCB-Prune-YOLO-P40-A8-Direct with ultralytics:
from ultralytics import YOLOvv8 model = YOLOvv8.from_pretrained("thangkt/PCB-Prune-YOLO-P40-A8-Direct") source = 'http://images.cocodataset.org/val2017/000000039769.jpg' model.predict(source=source, save=True) - Notebooks
- Google Colab
- Kaggle
Upload benchmark.json with huggingface_hub
Browse files- benchmark.json +6 -6
benchmark.json
CHANGED
|
@@ -6,12 +6,12 @@
|
|
| 6 |
"flops_estimate": 2242475200,
|
| 7 |
"gflops_estimate": 2.2424752,
|
| 8 |
"measurement_scope": "pure_model_forward_excludes_preprocess_and_nms",
|
| 9 |
-
"mean_latency_ms": 7.
|
| 10 |
-
"median_latency_ms": 7.
|
| 11 |
-
"p95_latency_ms": 8.
|
| 12 |
-
"fps":
|
| 13 |
-
"model": "outputs/finetune_direct/
|
| 14 |
-
"model_size_mb": 1.
|
| 15 |
"batch_size": 1,
|
| 16 |
"imgsz": 640,
|
| 17 |
"device": "cuda:0",
|
|
|
|
| 6 |
"flops_estimate": 2242475200,
|
| 7 |
"gflops_estimate": 2.2424752,
|
| 8 |
"measurement_scope": "pure_model_forward_excludes_preprocess_and_nms",
|
| 9 |
+
"mean_latency_ms": 7.624854440009585,
|
| 10 |
+
"median_latency_ms": 7.448529999237508,
|
| 11 |
+
"p95_latency_ms": 8.457600999463466,
|
| 12 |
+
"fps": 131.1500446162934,
|
| 13 |
+
"model": "outputs/finetune_direct/p40_a8_adamw_exact_ext/weights/best.pt",
|
| 14 |
+
"model_size_mb": 1.9655094146728516,
|
| 15 |
"batch_size": 1,
|
| 16 |
"imgsz": 640,
|
| 17 |
"device": "cuda:0",
|