JonneryR comments

Repositories
Issues
Comments

Results 3 comments of


                                            JonneryR

How to get quantized weights from QAT model?

@daverim Hi, I want to know how to get quantized biases. `biases_quantized = tf.math.round(layer.weights[1] / scale).numpy().astype(np.int32)` Thank you very much.

按操作文档finetune报错： styles = torch.LongTensor([[self.textnorm_int_dict[int(style)]] for style in text[:, 3]]).to(speech.device) IndexError: index 3 is out of bounds for dimension 1 with size 1

我也遇到了这个问题。他的代码看起来要对text去做padding，但我暂时没找到这个padding的代码在哪里，好像得自己补充了。

按操作文档finetune报错： styles = torch.LongTensor([[self.textnorm_int_dict[int(style)]] for style in text[:, 3]]).to(speech.device) IndexError: index 3 is out of bounds for dimension 1 with size 1

问题解决了，是dataset的选择问题，需要选择SenseVoiceCTCDataset，只有这里才有给text前面做padding的代码。