Wentao Xu
Wentao Xu
> ## ❓ Questions and Help > anyone can give me an introduction？I want to use HIST model on alpha158 Hi, to use the HIST model on alpha158, you need...
Yes, you can generate LSTM pre-train checkpoints and train with the same settings using LSTM. The training process is same as the steps you mentioned.
I also have a similar observation. The GAT has a much higher performance than the results reported in this paper. Could you share the code of baselines? or a part...
Hi, I use the code in Qlib for backtesting, referring to https://github.com/microsoft/qlib/tree/main/examples/benchmarks/HIST.
Hi, for the second question, you should indeed modify the label to close(T+2) / close(T+1) - 1 to enable the backtesting in Qlib because the Qlib only supports this label....
For the hyper-parameter d_model=64, these two versions of code have different labels, therefore they have different hyper-parameter d_model. For the hyper-parameter k, the Qlib version of HIST model uses k=50...
> Hello, I would like to ask why the hist model in qlib is much simplified and the market value is not used? Hi, because I want to simplify the...
> anyone can give me an introduction？I want to use HIST model on alpha158 Hi, to use the HIST model on alpha158, you need to change the encoder from GRU/LSTM...
The reply of @a919480698 is right.
Hi, there may be some errors in the paper's equations, please take the code as the correct standard.