# Confusion about feeding data in batches

**URL:** <https://discourse.edwardlib.org/t/confusion-about-feeding-data-in-batches/110>\
**Category:** General\
**Created:** [May 2, 2017, 6:37am UTC](https://discourse.edwardlib.org/t/confusion-about-feeding-data-in-batches/110 "2017-05-02T06:37:03Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![matttrent](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.edwardlib.org/matttrent/32/22_2.png) [@matttrent](https://discourse.edwardlib.org/u/matttrent)\
**Post date:** [May 2, 2017, 6:37am UTC](https://discourse.edwardlib.org/t/confusion-about-feeding-data-in-batches/110/1 "2017-05-02T06:37:03Z")

</div>

Hey all. Potentially dumb question that I wasn’t able to find an answer to in the docs/forum.

I’ve been understanding edward by means translating of the PyMC3 [tutorials](https://twiecki.github.io/blog/2016/06/01/bayesian-deep-learning/) into edward/tensorflow. And I’m a bit confused on how to do minibatch updates in edward.

Now, I’m confused on what I should be passing into the `data` parameter of `KLqp. __init__ ()` and what I should be passing into the `feed_dict` parameter of `inference.upate()`. My read of the documentation is that I pass in the minibatches as I have done below.

If I do that, do I need to pass anything to the `data` parameter? It didn’t seem to indicate that I needed to, but if I omitted them the model training didn’t seem to converge. And the training converges better if I pass in a `data` dict with the full data set in it (`X_train`/`y_train` vs the first batch).

Am I misunderstanding passing in data, or is there something more fundamental?

```auto
n_epoch = 500
n_iter_per_epoch = 50
batch_size = 10

inference = ed.KLqp({
        W_0: qW_0, b_0: qb_0, W_1: qW_1, b_1: qb_1, W_2: qW_2, b_2: qb_2
    }, 
    data={
        X: X_train[:batch_size, :],
        y: Y_train[:batch_size]
    }
)

inference.initialize(
    scale={y: float(N) / batch_size}
)
tf.global_variables_initializer().run()

losses = []
pbar = Progbar(n_epoch)
for epoch in range(n_epoch):
    loss = 0.0

    for t in range(n_iter_per_epoch):
        tmod = t % n_iter_per_epoch
        xt = X_train[tmod*batch_size:(tmod+1)*batch_size, :]
        yt = Y_train[tmod*batch_size:(tmod+1)*batch_size]

        info_dict = inference.update(feed_dict={
            X: xt, y: yt
        })
        loss += info_dict['loss']

    pbar.update(epoch+1, values={'loss': loss / float(n_iter_per_epoch)})

inference.finalize()

```

---

<div class="post-metadata">

**Author:** ![dustin](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.edwardlib.org/dustin/32/134_2.png) [@dustin](https://discourse.edwardlib.org/u/dustin)\
**Post date:** [May 2, 2017, 9:55am UTC](https://discourse.edwardlib.org/t/confusion-about-feeding-data-in-batches/110/2 "2017-05-02T09:55:52Z")

</div>

Batch training requires binding observed variables to TensorFlow placeholders in the `Inference` class. The intuition is that you are conditioning on values that you don’t specify until runtime.

In the `feed_dict` during `update()`, you then bind the TensorFlow placeholders to realized values (NumPy arrays).

The WIP tutorial [here](https://github.com/blei-lab/edward/pull/630) might be helpful.

---

<div class="post-metadata">

**Author:** ![matttrent](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.edwardlib.org/matttrent/32/22_2.png) [@matttrent](https://discourse.edwardlib.org/u/matttrent)\
**Post date:** [May 2, 2017, 9:07pm UTC](https://discourse.edwardlib.org/t/confusion-about-feeding-data-in-batches/110/3 "2017-05-02T21:07:44Z")

</div>

Aha, I see. I made it half-way there and instantiated `X` as a placeholder, but hadn’t done the same for `y` with an equivalent `y_ph`. Adding that, and plus passing `y_ph` instead of `y` in the `feed_dict` solved it. Thanks!

The [notebook](https://github.com/matttrent/mnist-uncertainty/blob/master/notebooks/edward_supervised_classification.ipynb) is very much a WIP itself, but if you think the latter parts would make an interesting contribution, I’d be happy to clean it up for PR.