owkin/FLamby

Wrong indentation at fed_ixi/benchmark line 144-154

Arctic-Xiangjian opened this issue · 3 comments

results = np.array(results)

Hello, I notice that in the above position, you make the result become np.array() which will make next different random seed experiment raise an error. that 'np.array() have no append'.

If delete the indentation of line 144-154. It will give the right output.

Here is the fix code.

        results.append(current_results_dict["client_test_0"])
    results = np.array(results)
    if log:
        for i in range(results.shape[0]):
            writer = SummaryWriter(log_dir=f"./runs/tests_seed{SEEDS[i]}")
            writer.add_scalar("DICE coefficient", results[i], 0)

    print("Benchmark Results on pooled IXI:")
    print(
        f"mDICE on {len(SEEDS)} runs: {results.mean():.2%} \\pm {results.std():.2%}"
    )

Thank you very much @Arctic-Xiangjian, we'll change that ASAP.

See #254

should be fixed by #254