Inicio Explorar Axuda
Iniciar sesión
kekezack
/
X_SSL_Net
0
0
Fork 0
Ficheiros Incidencias 0 Pull Requests 0 Wiki
Árbore: fbbbfe53c8
Ramas Etiquetas
X_SSL_Net
/
tmp
/
VMamba-main
/
analyze
/
mmpretrain_configs
/
configs
/
convmixer
/
metafile.yml

metafile.yml 2.3 KB
Histórico Raw

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061 
  1. Collections:
    2. 

  2.   - Name: ConvMixer
    3. 

  3.     Metadata:
    4. 

  4.       Training Data: ImageNet-1k
    5. 

  5.       Architecture:
    6. 

  6.         - 1x1 Convolution
    7. 

  7.         - LayerScale
    8. 

  8.     Paper:
    9. 

  9.       URL: https://arxiv.org/abs/2201.09792
    10. 

  10.       Title: Patches Are All You Need?
    11. 

  11.     README: configs/convmixer/README.md
    12. 

  12. 

  13. Models:
    14. 

  14.   - Name: convmixer-768-32_3rdparty_in1k
    15. 

  15.     Metadata:
    16. 

  16.       FLOPs: 19623051264
    17. 

  17.       Parameters: 21110248
    18. 

  18.     In Collection: ConvMixer
    19. 

  19.     Results:
    20. 

  20.       - Dataset: ImageNet-1k
    21. 

  21.         Metrics:
    22. 

  22.           Top 1 Accuracy: 80.16
    23. 

  23.           Top 5 Accuracy: 95.08
    24. 

  24.         Task: Image Classification
    25. 

  25.     Weights: https://download.openmmlab.com/mmclassification/v0/convmixer/convmixer-768-32_3rdparty_10xb64_in1k_20220323-bca1f7b8.pth
    26. 

  26.     Config: configs/convmixer/convmixer-768-32_10xb64_in1k.py
    27. 

  27.     Converted From:
    28. 

  28.       Weights: https://github.com/tmp-iclr/convmixer/releases/download/v1.0/convmixer_768_32_ks7_p7_relu.pth.tar
    29. 

  29.       Code: https://github.com/locuslab/convmixer
    30. 

  30.   - Name: convmixer-1024-20_3rdparty_in1k
    31. 

  31.     Metadata:
    32. 

  32.       FLOPs: 5550112768
    33. 

  33.       Parameters: 24383464
    34. 

  34.     In Collection: ConvMixer
    35. 

  35.     Results:
    36. 

  36.       - Dataset: ImageNet-1k
    37. 

  37.         Metrics:
    38. 

  38.           Top 1 Accuracy: 76.94
    39. 

  39.           Top 5 Accuracy: 93.36
    40. 

  40.         Task: Image Classification
    41. 

  41.     Weights: https://download.openmmlab.com/mmclassification/v0/convmixer/convmixer-1024-20_3rdparty_10xb64_in1k_20220323-48f8aeba.pth
    42. 

  42.     Config: configs/convmixer/convmixer-1024-20_10xb64_in1k.py
    43. 

  43.     Converted From:
    44. 

  44.       Weights: https://github.com/tmp-iclr/convmixer/releases/download/v1.0/convmixer_1024_20_ks9_p14.pth.tar
    45. 

  45.       Code: https://github.com/locuslab/convmixer
    46. 

  46.   - Name: convmixer-1536-20_3rdparty_in1k
    47. 

  47.     Metadata:
    48. 

  48.       FLOPs: 48713170944
    49. 

  49.       Parameters: 51625960
    50. 

  50.     In Collection: ConvMixer
    51. 

  51.     Results:
    52. 

  52.       - Dataset: ImageNet-1k
    53. 

  53.         Metrics:
    54. 

  54.           Top 1 Accuracy: 81.37
    55. 

  55.           Top 5 Accuracy: 95.61
    56. 

  56.         Task: Image Classification
    57. 

  57.     Weights: https://download.openmmlab.com/mmclassification/v0/convmixer/convmixer-1536_20_3rdparty_10xb64_in1k_20220323-ea5786f3.pth
    58. 

  58.     Config: configs/convmixer/convmixer-1536-20_10xb64_in1k.py
    59. 

  59.     Converted From:
    60. 

  60.       Weights: https://github.com/tmp-iclr/convmixer/releases/download/v1.0/convmixer_1536_20_ks9_p7.pth.tar
    61. 

  61.       Code: https://github.com/locuslab/convmixer
    62.