YonV: Don't open Issues like this 不建议像这样开问题 → Found dtype Double but expected Float about elegantrl HOT 3 CLOSED

YonV: Don't open Issues like this 不建议像这样开问题 → Found dtype Double but expected Float

The error message has made it very clear: it needs to accept float instead of double data. Then use the code to convert the data format. I don't recommended to open an issue for it.
报错信息已经非常清楚地表明：它需要接受 float 而非 double 的数据。那么使用代码对数据格式进行转换就解决了。这种问题不建议占用 Issue 的版面寻求解决。

array = np.array(array, dtype=np.floate32)
array = array.astype(np.float32)  

from elegantrl.

好的，谢谢！

from elegantrl.

critic_loss = self.criterion(q1_value, q_target.type_as(q1_value)) + self.criterion(q2_value, q_target.type_as(q1_value))
解决了谢谢 这个地方强转一下就行了 慢慢摸索还是能行的

from elegantrl.