# Shape mismatch error for out of sample inference

**URL:** <https://discourse.pymc.io/t/shape-mismatch-error-for-out-of-sample-inference/15089>\
**Category:** v5\
**Tags:** modeling\
**Created:** [September 18, 2024, 8:46am UTC](https://discourse.pymc.io/t/shape-mismatch-error-for-out-of-sample-inference/15089 "2024-09-18T08:46:47Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![Max1](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/max1/32/8514_2.png) [@Max1](https://discourse.pymc.io/u/Max1)\
**Post date:** [September 18, 2024, 8:46am UTC](https://discourse.pymc.io/t/shape-mismatch-error-for-out-of-sample-inference/15089/1 "2024-09-18T08:46:47Z")

</div>

Hello, I know there’s a few questions on this topic and I (believe!) I have tried all the suggested solutions but none work for me.

I’m running a simple Bayesian linear model, and want to perform inference on a 2d (data frame) that has a different number of rows to my training data. This doesn’t work - if they have the same shape it’s all fine, but as soon as I change the number of rows, well …

In my opinion I have tried everything, coords, mutable coords, adding the observed values for the sampling as a Data argument and re-setting those during inference (even though I don’t know these values of course!) - but nothing works. Any help or pointers massively appreciated, I feel this shouldn’t be so difficult?!

Minimal working example:

```auto
covariates = pd.DataFrame(np.random.randint(1,5,(38, 9)), columns=[f'feature_{i}' for i in range(9)])

coords = {'obs': covariates.index, 'features':covariates.columns}
coords_mutable = {'obs': np.arange(len(covariates))}

with pm.Model(coords=coords, coords_mutable=coords_mutable) as base_model:

    covars_data = pm.Data("covars_data", covariates, dims=['obs', 'features'])

    mu = pm.Normal("mu", 0, sigma=1)
    sigma = pm.HalfCauchy("sigma", beta=10)
    n = covariates.shape[0]
    n_cov = covariates.shape[1]
    
    intercept = pm.Normal("intercept", mu=mu, sigma=sigma, shape=n)

    beta = pm.Normal("beta", mu=0, sigma=1, shape=n_cov, dims="features")
    mean_val = intercept + covars_data @ beta

    data = pd.DataFrame(np.random.randint(1,5,(38, 5)), columns=['a','b','c','d','e'])

    endpt1_data = pm.Data(
        "endpt1_data", data.a
    )
    se1_data = pm.Data("se1_data", data.b)

    sigma_data = pm.Data("sigma_data", data.c, dims='obs')
    obs2_data = pm.Data('obs2_data', data.e, dims='obs')

    # Meta-analytic likelihood of endpoint 2
    est2_dist = pm.Normal(
        "endpt2",
        mu=mean_val,
        sigma=sigma_data,
        observed=obs2_data,
        dims='obs',
    )

    idata = pm.sample(100)

### prediction
covariates = pd.DataFrame(np.random.randint(1,5,(1, 9)), columns=[f'feature_{i}' for i in range(9)])

with base_model:
            
    pm.set_data({"covars_data": covariates}, coords={'obs': np.arange(len(covariates))})
    # pm.set_data({"obs2_data": [0]}, coords={'obs': np.arange(len(covariates))}) # this seems nonsense?? 
    # pm.set_data({"sigma_data": [1]}, coords={'obs': np.arange(len(covariates))}) 

    ppc = pm.sample_posterior_predictive(idata)

```

If I just delete `est2_dist` from the model training this runs, and even if I keep it and just delete observed= it still runs … so that seems to be the issue!? But then how do I solve this?

And error is:  
`ValueError: Shape mismatch: A.shape[0] != y.shape[0] Apply node that caused the error: CGemv{no_inplace}(intercept, 1.0, covars_data, beta, 1.0) Toposort index: 0 Inputs types: [TensorType(float64, shape=(38,)), TensorType(float64, shape=()), TensorType(float64, shape=(None, None)), TensorType(float64, shape=(9,)), TensorType(float64, shape=())] Inputs shapes: [(38,), (), (1, 9), (9,), ()] Inputs strides: [(8,), (), (72, 8), (8,), ()] Inputs values: ['not shown', array(1.), 'not shown', 'not shown', array(1.)]`

Thank you!

---

<div class="post-metadata">

**Author:** ![ricardoV94](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/ricardov94/32/5775_2.png) [@ricardoV94](https://discourse.pymc.io/u/ricardoV94)\
**Post date:** [September 18, 2024, 1:37pm UTC](https://discourse.pymc.io/t/shape-mismatch-error-for-out-of-sample-inference/15089/3 "2024-09-18T13:37:37Z")

</div>

Are you sure your intercept should have `shape=n`?

---

<div class="post-metadata">

**Author:** ![Max1](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/max1/32/8514_2.png) [@Max1](https://discourse.pymc.io/u/Max1)\
**Post date:** [September 18, 2024, 2:25pm UTC](https://discourse.pymc.io/t/shape-mismatch-error-for-out-of-sample-inference/15089/4 "2024-09-18T14:25:51Z")

</div>

wow … that may have fixed it - are you able to explain why? Agreed I only need a single intercept, but just for my pymc understanding, why would this have broken anything?

---

<div class="post-metadata">

**Author:** ![ricardoV94](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/ricardov94/32/5775_2.png) [@ricardoV94](https://discourse.pymc.io/u/ricardoV94)\
**Post date:** [September 18, 2024, 2:34pm UTC](https://discourse.pymc.io/t/shape-mismatch-error-for-out-of-sample-inference/15089/5 "2024-09-18T14:34:08Z")

</div>

The reason it broke is that you defined it with a static shape, based on `covariates`, instead of a dynamic one based on `covars_data`, so PyMC wouldn’t resize it after you updated the MutableData.

But it wouldn’t make sense to have a RV size change between sample and posterior predictive, as it would be unclear how to reuse the psoterior inferences (pymc would end up sampling the posterior predictive of that RV as well, which in you case would be a prior draw from the distribution)

---

<div class="post-metadata">

**Author:** ![Max1](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/max1/32/8514_2.png) [@Max1](https://discourse.pymc.io/u/Max1)\
**Post date:** [September 20, 2024, 8:11am UTC](https://discourse.pymc.io/t/shape-mismatch-error-for-out-of-sample-inference/15089/6 "2024-09-20T08:11:55Z")

</div>

Sorry I have one further question. Weirdly, this now only works for me if I set obs2\_data and sigma\_data to some dummy values … I read that in some other post [here](https://discourse.pymc.io/t/predict-with-new-coords-leads-to-conflicting-sizes/10591/2), but that cannot be correct, right?! Cause how would it then differentiate between these and actual variables I have observed and need to update (such as covars\_data here)? Appreciate the help!