# Optimizing for FNC and TB scores

**URL:** <https://forum.numer.ai/t/optimizing-for-fnc-and-tb-scores/5132>\
**Category:** Tournament\
**Created:** [March 22, 2022, 10:21pm UTC](https://forum.numer.ai/t/optimizing-for-fnc-and-tb-scores/5132 "2022-03-22T22:21:59Z")\
**Posts on this page:** 1\
**Showing post:** 32

<div class="post-metadata">

**Author:** ![olivepossum](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/olivepossum/32/2259_2.png) [@olivepossum](https://forum.numer.ai/u/olivepossum)\
**Post date:** [May 26, 2022, 9:42pm UTC](https://forum.numer.ai/t/optimizing-for-fnc-and-tb-scores/5132/32 "2022-05-26T21:42:01Z")

</div>

Think I came up with an implementation that would work on GPU as uses PyTorch. However, reading this post [True Contribution Details](http://forum.numer.ai/t/true-contribution-details/5128), exposure dissimilarity seems to be relevant just combined with FNCv3 on a multiplicative way so it might not make sense to use it without it.  
Any feedback is more than welcome!

```auto
for f in feature_cols:
  train_data[f] -= 0.5

for epoch in range(epochs):
    np.random.shuffle(era_list)
    batch_count = 0
    acc_loss_train = 0
    for era in era_list:
        batch_count += 1

        # get features and target from data and put in tensors
        features = torch.tensor(train_data[train_data.erano == era].filter(like='feature').values)
        target = torch.tensor(train_data[train_data.erano == era]['target'])

        # zero gradient buffer and get model output
        optimizer.zero_grad()
        model.train()
        model_output = model(features)

        orig_loss = -numerair_tb(model_output, target)

        #dissimilarity
        train_era = train_data[train_data.erano == era]

        example_preds = torch.as_tensor(train_era['example_preds'].values) #Needs to be created previously
        example_preds = example_preds - example_preds.mean()
        corr_example_preds = (features.T * example_preds).sum(dim=1) / ((features.T * features.T).sum(dim=1) * (example_preds * example_preds).sum()).sqrt()

        preds = model_output
        preds = preds - preds.mean()
        corr_preds = (features.T * preds).sum(dim=1) / ((features.T * features.T).sum(dim=1) * (preds * preds).sum()).sqrt()

        num = corr_preds.pinverse(rcond=1e-6).dot(corr_example_preds)
        denom = corr_example_preds.pinverse(rcond=1e-6).dot(corr_example_preds)

        dissimilarity = (num/denom).sum()

        #final loss
        loss = - orig_loss + dissimilarity

        acc_loss_train += loss 
        loss.backward()
        optimizer.step()

    loss_train = acc_loss_train / batch_count

```

---

_[View the full topic](https://forum.numer.ai/t/optimizing-for-fnc-and-tb-scores/5132)._