Add filesystem & use it in one embedding

[mosout]

[github-actions[bot]]

Speed stats:

GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.8ms (= 12984.3ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 143.6ms (= 14362.9ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.11 (= 143.6ms / 129.8ms)

OneFlow resnet50 time: 76.2ms (= 7621.4ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 87.6ms (= 8762.9ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.15 (= 87.6ms / 76.2ms)

OneFlow resnet50 time: 50.1ms (= 10011.3ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 60.9ms (= 12177.8ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.22 (= 60.9ms / 50.1ms)

OneFlow resnet50 time: 37.3ms (= 7457.5ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 45.5ms (= 9105.7ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.22 (= 45.5ms / 37.3ms)

OneFlow resnet50 time: 31.8ms (= 6359.2ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 42.5ms (= 8505.7ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.34 (= 42.5ms / 31.8ms)

OneFlow swin dataloader time: 0.260s (= 51.991s / 200, num_workers=1)
PyTorch swin dataloader time: 0.154s (= 30.717s / 200, num_workers=1)
Relative speed: 0.591 (= 0.154s / 0.260s)

OneFlow swin dataloader time: 0.072s (= 14.339s / 200, num_workers=4)
PyTorch swin dataloader time: 0.042s (= 8.424s / 200, num_workers=4)
Relative speed: 0.587 (= 0.042s / 0.072s)

OneFlow swin dataloader time: 0.043s (= 8.620s / 200, num_workers=8)
PyTorch swin dataloader time: 0.023s (= 4.549s / 200, num_workers=8)
Relative speed: 0.528 (= 0.023s / 0.043s)

❌ OneFlow resnet50 time: 145.9ms (= 14586.0ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 171.0ms (= 17104.2ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.17 (= 171.0ms / 145.9ms)

OneFlow resnet50 time: 94.9ms (= 9487.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 113.2ms (= 11321.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.19 (= 113.2ms / 94.9ms)

OneFlow resnet50 time: 67.1ms (= 13426.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 87.8ms (= 17554.7ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.31 (= 87.8ms / 67.1ms)

OneFlow resnet50 time: 55.9ms (= 11173.8ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.3ms (= 14858.7ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.33 (= 74.3ms / 55.9ms)

OneFlow resnet50 time: 48.2ms (= 9636.9ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 69.8ms (= 13968.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.45 (= 69.8ms / 48.2ms)

[github-actions[bot]]

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/8800/