TypeError: add got incompatible shapes for broadcasting: (58,), (54,).

[MarkusStefan]

---

TypeError Traceback (most recent call last)
in <cell line: 2>()
4 seed=SEED)
5 else:
----> 6 new_predictions = mmm.predict(media=media_scaler.transform(media_data_test),
7 extra_features=extra_features_scaler.transform(extra_features_test),
8 seed=SEED)

17 frames
/usr/local/lib/python3.10/dist-packages/lightweight_mmm/lightweight_mmm.py in predict(self, media, extra_features, media_gap, target_scaler, seed)
518 if seed is None:
519 seed = utils.get_time_seed()
--> 520 prediction = self._predict(
521 rng_key=jax.random.PRNGKey(seed=seed),
522 media_data=full_media,

[... skipping hidden 12 frame]

/usr/local/lib/python3.10/dist-packages/lightweight_mmm/lightweight_mmm.py in _predict(self, rng_key, media_data, extra_features, media_prior, degrees_seasonality, frequency, transform_function, weekday_seasonality, model, posterior_samples, custom_priors)
441 The predictions for the given data.
442 """
--> 443 return infer.Predictive(
444 model=model, posterior_samples=posterior_samples)(
445 rng_key=rng_key,

/usr/local/lib/python3.10/dist-packages/numpyro/infer/util.py in call(self, rng_key, *args, **kwargs)
1009 """
1010 if self.batch_ndims == 0 or self.params == {} or self.guide is None:
-> 1011 return self._call_with_params(rng_key, self.params, args, kwargs)
1012 elif self.batch_ndims == 1: # batch over parameters
1013 batch_size = jnp.shape(tree_flatten(self.params)[0][0])[0]

/usr/local/lib/python3.10/dist-packages/numpyro/infer/util.py in _call_with_params(self, rng_key, params, args, kwargs)
986 )
987 model = substitute(self.model, self.params)
--> 988 return _predictive(
989 rng_key,
990 model,

/usr/local/lib/python3.10/dist-packages/numpyro/infer/util.py in _predictive(rng_key, model, posterior_samples, batch_shape, return_sites, infer_discrete, parallel, model_args, model_kwargs)
823 rng_key = rng_key.reshape(batch_shape + key_shape)
824 chunk_size = num_samples if parallel else 1
--> 825 return soft_vmap(
826 single_prediction, (rng_key, posterior_samples), len(batch_shape), chunk_size
827 )

/usr/local/lib/python3.10/dist-packages/numpyro/util.py in soft_vmap(fn, xs, batch_ndims, chunk_size)
417 fn = vmap(fn)
418
--> 419 ys = lax.map(fn, xs) if num_chunks > 1 else fn(xs)
420 map_ndims = int(num_chunks > 1) + int(chunk_size > 1)
421 ys = tree_map(

[... skipping hidden 12 frame]

/usr/local/lib/python3.10/dist-packages/numpyro/infer/util.py in single_prediction(val)
796 )
797 else:
--> 798 model_trace = trace(
799 seed(substitute(masked_model, samples), rng_key)
800 ).get_trace(*model_args, **model_kwargs)

/usr/local/lib/python3.10/dist-packages/numpyro/handlers.py in get_trace(self, *args, **kwargs)
169 :return: OrderedDict containing the execution trace.
170 """
--> 171 self(*args, **kwargs)
172 return self.trace
173

/usr/local/lib/python3.10/dist-packages/numpyro/primitives.py in call(self, *args, **kwargs)
103 return self
104 with self:
--> 105 return self.fn(*args, **kwargs)
106
107

/usr/local/lib/python3.10/dist-packages/numpyro/primitives.py in call(self, *args, **kwargs)
103 return self
104 with self:
--> 105 return self.fn(*args, **kwargs)
106
107

/usr/local/lib/python3.10/dist-packages/numpyro/primitives.py in call(self, *args, **kwargs)
103 return self
104 with self:
--> 105 return self.fn(*args, **kwargs)
106
107

/usr/local/lib/python3.10/dist-packages/numpyro/primitives.py in call(self, *args, **kwargs)
103 return self
104 with self:
--> 105 return self.fn(*args, **kwargs)
106
107

/usr/local/lib/python3.10/dist-packages/numpyro/primitives.py in call(self, *args, **kwargs)
103 return self
104 with self:
--> 105 return self.fn(*args, **kwargs)
106
107

/usr/local/lib/python3.10/dist-packages/lightweight_mmm/models.py in media_mix_model(media_data, target_data, media_prior, degrees_seasonality, frequency, transform_function, custom_priors, transform_kwargs, weekday_seasonality, extra_features)
410 # expo_trend is B(1, 1) so that the exponent on time is in [.5, 1.5].
411 prediction = (
--> 412 intercept + coef_trend * trend ** expo_trend +
413 seasonality * coef_seasonality +
414 jnp.einsum(media_einsum, media_transformed, coef_media))

/usr/local/lib/python3.10/dist-packages/jax/_src/numpy/array_methods.py in op(self, *args)
741 def forward_operator_to_aval(name):
742 def op(self, *args):
--> 743 return getattr(self.aval, f"{name}")(self, *args)
744 return op
745

/usr/local/lib/python3.10/dist-packages/jax/_src/numpy/array_methods.py in deferring_binary_op(self, other)
269 args = (other, self) if swap else (self, other)
270 if isinstance(other, _accepted_binop_types):
--> 271 return binary_op(*args)
272 # Note: don't use isinstance here, because we don't want to raise for
273 # subclasses, e.g. NamedTuple objects that may override operators.

[... skipping hidden 12 frame]

/usr/local/lib/python3.10/dist-packages/jax/src/numpy/ufuncs.py in fn(x1, x2)
97 def fn(x1, x2, /):
98 x1, x2 = promote_args(numpy_fn.name, x1, x2)
---> 99 return lax_fn(x1, x2) if x1.dtype != np.bool else bool_lax_fn(x1, x2)
100 fn.qualname = f"jax.numpy.{numpy_fn.name}"
101 fn = jit(fn, inline=True)

[... skipping hidden 7 frame]

/usr/local/lib/python3.10/dist-packages/jax/_src/lax/lax.py in broadcasting_shape_rule(name, *avals)
1597 result_shape.append(non_1s[0])
1598 else:
-> 1599 raise TypeError(f'{name} got incompatible shapes for broadcasting: '
1600 f'{", ".join(map(str, map(tuple, shapes)))}.')
1601

TypeError: add got incompatible shapes for broadcasting: (58,), (54,).

[Pavantejapenugonda]

Even i am getting the issue, looking for the solution for it

[MarkusStefan]

Installing an older version of numpyro resolved my issue !pip numpyro == 0.13.2

[masifkingpin]

I had the same problem and 0.13.2 version of numpyro was not working for me so I used the following command to install numpyro while installing mmm, matplotlib etc:

!pip install numpyro==0.13.1

[shivahari15091994]

I am also facing the same problem. Appreciate if anyone has solution for this. Thanks

[MarkusStefan]

just install an older version of numpyro as stated in the comments above

[jingwg]

When i install an older version of numpyro, I have following issues with import . Any idea how to solve this?

ModuleNotFoundError Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_12544\2990551649.py in
1 import pandas as pd
----> 2 from lightweight_mmm import preprocessing, lightweight_mmm, plot, optimize_media
3 import jax.numpy as jnp
4 from sklearn.metrics import mean_absolute_percentage_error

~\Anaconda3\envs\python3\lib\site-packages\lightweight_mmm\preprocessing.py in
24 from statsmodels.stats.outliers_influence import variance_inflation_factor
25 from statsmodels.tools.tools import add_constant
---> 26 from lightweight_mmm.core import core_utils
27
28

~\Anaconda3\envs\python3\lib\site-packages\lightweight_mmm\core\core_utils.py in
20 import jax.numpy as jnp
21
---> 22 from numpyro import distributions as dist
23
24 # pylint: disable=g-import-not-at-top

~\Anaconda3\envs\python3\lib\site-packages\numpyro_init_.py in
4 import logging
5
----> 6 from numpyro import compat, diagnostics, distributions, handlers, infer, ops, optim
7 from numpyro.distributions.distribution import enable_validation, validation_enabled
8 from numpyro.infer.inspect import render_model

~\Anaconda3\envs\python3\lib\site-packages\numpyro\infer_init_.py in
3
4 from numpyro.infer.barker import BarkerMH
----> 5 from numpyro.infer.elbo import (
6 ELBO,
7 RenyiELBO,

~\Anaconda3\envs\python3\lib\site-packages\numpyro\infer\elbo.py in
23 log_density,
24 )
---> 25 from numpyro.ops.provenance import eval_provenance
26 from numpyro.util import _validate_model, check_model_guide_match, find_stack_level
27

~\Anaconda3\envs\python3\lib\site-packages\numpyro\ops\provenance.py in
6 import jax.core as core
7 from jax.experimental.pjit import pjit_p
----> 8 import jax.extend.linear_util as lu
9 from jax.interpreters.partial_eval import trace_to_jaxpr_dynamic
10 from jax.interpreters.pxla import xla_pmap_p

ModuleNotFoundError: No module named 'jax.extend.linear_util'

[AkiroSR]

install an older version of jax.
'jax.extend.linear_util' was removed in jax after 0.4.23 (currently in 0.4.25)

[fehiepsi]

Sorry for the breakage. Could you try

pip install --upgrade git+https://github.com/google/lightweight_mmm.git