INNER CODE UNIT · Python

forward_returns

initial-d/ml-quant-trading · src/mlquant/cli/main.py:271

        forward_returns=fwd_ret[: oos + 1],
    )
    n_features = ds.features.shape[1]
    model = MLPRegressor(in_dim=n_features, hidden=cfg.get("hidden", 128))
    trainer = Trainer(model, AdjMSELoss(gamma=cfg.get("loss_gamma", 0.1)),
                      TrainConfig(**(cfg.train.to_dict() if "train" in cfg else {})))
    trainer.fit(ds)

    # Write predictions for every tradable out-of-sample (t, n) cell.
    with torch.no_grad():
        T, N, F = feat_blob["factors"].shape
        flat = feat_blob["factors"].reshape(-1, F)
        pred = model(flat.to(trainer.cfg.device)).cpu().reshape(T, N)
    pred = pred * feat_blob["mask"].float()
    pred[:oos] = 0.0
    torch.save({"pred": pred, "oos_start": oos}, art / "predictions.pt")
    click.echo(
        f"wrote predictions: {pred.shape}  (trained on dates [0, {oos}),"

View source record →