metafile.yml 2.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778
  1. Collections:
  2. - Name: SSD
  3. Metadata:
  4. Training Data: COCO
  5. Training Techniques:
  6. - SGD with Momentum
  7. - Weight Decay
  8. Training Resources: 8x V100 GPUs
  9. Architecture:
  10. - VGG
  11. Paper:
  12. URL: https://arxiv.org/abs/1512.02325
  13. Title: 'SSD: Single Shot MultiBox Detector'
  14. README: configs/ssd/README.md
  15. Code:
  16. URL: https://github.com/open-mmlab/mmdetection/blob/v2.14.0/mmdet/models/dense_heads/ssd_head.py#L16
  17. Version: v2.14.0
  18. Models:
  19. - Name: ssd300_coco
  20. In Collection: SSD
  21. Config: configs/ssd/ssd300_coco.py
  22. Metadata:
  23. Training Memory (GB): 9.9
  24. inference time (ms/im):
  25. - value: 22.88
  26. hardware: V100
  27. backend: PyTorch
  28. batch size: 1
  29. mode: FP32
  30. resolution: (300, 300)
  31. Epochs: 120
  32. Results:
  33. - Task: Object Detection
  34. Dataset: COCO
  35. Metrics:
  36. box AP: 25.5
  37. Weights: https://download.openmmlab.com/mmdetection/v2.0/ssd/ssd300_coco/ssd300_coco_20210803_015428-d231a06e.pth
  38. - Name: ssd512_coco
  39. In Collection: SSD
  40. Config: configs/ssd/ssd512_coco.py
  41. Metadata:
  42. Training Memory (GB): 19.4
  43. inference time (ms/im):
  44. - value: 32.57
  45. hardware: V100
  46. backend: PyTorch
  47. batch size: 1
  48. mode: FP32
  49. resolution: (512, 512)
  50. Epochs: 120
  51. Results:
  52. - Task: Object Detection
  53. Dataset: COCO
  54. Metrics:
  55. box AP: 29.5
  56. Weights: https://download.openmmlab.com/mmdetection/v2.0/ssd/ssd512_coco/ssd512_coco_20210803_022849-0a47a1ca.pth
  57. - Name: ssdlite_mobilenetv2-scratch_8xb24-600e_coco
  58. In Collection: SSD
  59. Config: configs/ssd/ssdlite_mobilenetv2-scratch_8xb24-600e_coco.py
  60. Metadata:
  61. Training Memory (GB): 4.0
  62. inference time (ms/im):
  63. - value: 14.3
  64. hardware: V100
  65. backend: PyTorch
  66. batch size: 1
  67. mode: FP32
  68. resolution: (320, 320)
  69. Epochs: 600
  70. Results:
  71. - Task: Object Detection
  72. Dataset: COCO
  73. Metrics:
  74. box AP: 21.3
  75. Weights: https://download.openmmlab.com/mmdetection/v2.0/ssd/ssdlite_mobilenetv2_scratch_600e_coco/ssdlite_mobilenetv2_scratch_600e_coco_20210629_110627-974d9307.pth