Friday, February 08, 2019
Last week, we released nnabla packages v1.0.11! The main features updated from the previous version are described below:
nnabla.models API which allows users to use state-of-the-art pretrained models for both inference and training, without having to manually train the model from scratch, as shown below:
from nnabla.models.imagenet import ResNet
model = ResNet(18)
batch_size = 1
x = nn.Variable((batch_size,) + model.input_shape)
y = model(x, training=False)
Following models pretrained on ImageNet are available;
Add functions: IsInf, IsNaN, ResetNaN, ResetInf, and Where
Add clear_buffer flag to forward_all
Add 3D support for pooling functions.
[Experimental] PyTorch-like Functions
Serialization of SolverState
Add CuDNN max and average pooling for 3D case.
Use dedicated function to determine workspace size for alogorithm.
Improve onnx exporter
improve onnx import
[Experimental] Trainer API
On memory api
Support multiple dataset in .proto
Added learing rate scheduler
Add graph converters for inference
Efficient Neural Architecture Search (ENAS)