Skip to main content

oneflow tools

Project description

OneUtils

A collection of useful tools for oneflow

Installation

pip install one-utils==0.0.2

Usage

  • Weight Transfer
from one_utils import convert_torch_to_flow
model = model()
torch_weight = "path/to/pytorch_pretrained_weight.pth"
save_path = "path/to/save"
convert_torch_to_flow(model, torch_weight_path, save_path)
  • Eval Pytorch Model Acc on ImageNet
from one_utils import eval_torch_acc
model = model()
data_path = "path/to/imagenet/val"
eval_torch_acc(model, data_path, n_gpu_use=1)
  • Eval OneFlow Model Acc on ImageNet
from one_utils import eval_flow_acc
import multiprocessing as mp
mp.set_start_method("spawn")
model = model()
data_path = "path/to/imagenet/val"
eval_flow_acc(model, data_path)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

one-utils-0.0.3.tar.gz (3.6 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page