-
Now that inception v4 pertained network weights is out. https://research.googleblog.com/2016/08/improving-inception-and-image.html
Is there any plan to support loading this pretrained inception v4…
-
## Mobile inference frameworks benchmark (4*ARM_CPU)
Network|VOC mAP(0.5)|COCO mAP(0.5)|Resolution|Inference time (NCNN/Kirin 990)|Inference time (NNN arm82/Kirin 990)|FLOPS|Weight size
:---:|:---:|…
-
Prior to filing: check that this should be a bug instead of a feature request. Everything supported, including the compatible versions of TensorFlow, is listed in the overview page of each technique. …
-
Hello, I transcoded a PyTorch HRNet model to uint8 and obtained the following output in the ```prerun_graph_multithread``` function. The source code is similar to ```tm_unet.cpp``` adapted for uint8 b…
pl561 updated
3 years ago
-
Hello,
I trained a SPLADE model on my own recently. To reduce the inference time, I tried to export my model to ONNX with `torch.onnx.export()` but I encountered a few errors.
Is there a tutori…
ntnq4 updated
9 months ago
-
## 🚀 Feature
When we apply some float-point operation by a constant tensor, we may need to check if the input tensor is also float.
## Motivation
To avoid bugs, as described here, which appea…
-
### Feature request
Added `validate_checkpoint_key: bool` in `from_pretrained`
### Motivation
When pytorch loads state dict there is an option `strict=True` but there is no counterpart in `from_pre…
-
These are increasing in popularity, and alongside the lack of documentation for loading local models (#14) , makes it difficult to use these models. My current workflow is:
- Download the CKPT mode…
-
Thanks to the author for his wonderful research. I am currently reproducing the performance of the model on the NGSIM data set, but the results of my training are larger than those shown in your paper…
-
/home/mengfei/anaconda3/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
return f(*args, **kwds)
/home/…