Jimes
Jimes
> use original model with strict=False I mean it is different with the official interface about the Quantization Aware Training(https://pytorch.org/docs/stable/quantization.html),I don't know how to get the int8 model from the...
> not sure what int8 model means. > > anyway, i use this method. at model init part, > > during training self.conv1 = lsqconv(~~) > > during inferencing self.conv1...
pytorch1.8.1 , gcc 7.5.0 > Hello, please give more more info: your torch version, building environment, etc.
> Could you plrease provide a configuration file for the model evaluation and the related code? I have found builders such as flickr, textvqa, nocaps, as well as the corresponding...
大佬，想问问rt-detr的Efficient Hybrid Encoder模块设计的motivation是？比如CCFM的结构设计主要是参考哪篇论文或怎么想出来的，还有怎么发现用单层的S5进行seft_att比多个尺度的特征进行seft_att效果好。期待大佬的解答 @lyuwenyu
Thanks! > From backbone VideoLLaMA
> The relative improvement is the absolute value of (our score - previous score) / previous score. thanks, I see.