Alxendx

So.....Does it enable me to convert PyTorch model (NCHW) to TFlite (NHWC)????

I have another question, does tensorflow2.3 support folding op of conv+BN+Relu or conv+BN in QAT like pytorch? My model contains a lot of conv+BN+relu, but when i use `quant_aware_model =...

i faced the the same problem, thank you for your answer.

Could anyone have solved this problem?

加这个anchor，对训练阶段有什么帮助吗？

> 你好，目前是初始化训练， 最近没时间修改了， 你可以自己修改一下，难度不高的 请问下，mobileNet-yolov3要从哪里获得coco的预训练权重呢？

> I did not train MAT model, I fixed the issue of fp16 overflow in the inference code. The related modifications are [7e028c3](https://github.com/Sanster/lama-cleaner/commit/7e028c39082cef45fcb8ce23cb871c121c763dd6) and the following two commits. Thanks. Is...

> You don't need to train to obtain fp16 weights, just load the weights and convert them to fp16 before saving. I modified MAT original project according to your modification...

> I have download "aspose-words-20.12-jdk17-cracked.jar" and "javassist-3.24.1-GA.jar" but I can't run it > It always crash: > > ``` > java.lang.NoClassDefFoundError: Failed resolution of: Lcom/aspose/words/internal/zzYVE; > at com.aspose.words.internal.zz3E.zzZ(Unknown Source:243) >...

> I had to revert back to Keras-1.2, no other option except for altering the entire code. I have back to Keras-1.2, but the question still exists......QAQ