Jacob Burnim

You can use `joint.sample_distributions()` to return both samples and an instance of each component distribution (conditioned on the samples from the distributions on which it depends), and then can call...

In this case, `joint.sample_distributions(num_samples)` is roughly equivalent to: ```python distributions = {} samples = {} distributions['e'] = tfd.Exponential([100, 200]) samples['e'] = distributions['e'].sample(num_samples) distributions['n'] = tfd.Normal(loc=0., scale=2.) samples['n'] = distributions['n'].sample(num_samples) distributions['g']...

Good catch -- there are two typos (extra commas at the end of lines `distributions['m'] = ...` and `distributions['x'] = ...`), and, to get the correct shapes, `num_samples` should only...

NOTE: TFP 0.22.0 has been released.

Thanks for the report! I believe we have changed all of our documentation/examples in the TFP codebase to use `tf_keras` instead of `tf.keras`. But it doesn't look like these changes...

Does the error go away if you add: ```python import tf_keras ``` and replace `tf.optimizers` with `tf_keras.optimizers`? In TF 2.16+, `tf.optimizers` refers to the Keras 3 version of the optimizer,...

From the TFP 0.24.0 release notes -- https://github.com/tensorflow/probability/releases/tag/v0.24.0 : > NOTE: In TensorFlow 2.16+, `tf.keras` (and `tf.initializers`, `tf.losses`, and `tf.optimizers`) refers to Keras 3. TensorFlow Probability is not compatible with...

What version of g++ are you using? In particular, what is the output of running `g++ --version`? It looks like gcc6 now treats "narrowing conversions" as an error rather than...

> this error came out, so I edited the location of yices_c.h in yices_solver.cc to my yices directory. then Have you set `YICES_DIR` in `crest/src/Makefile` to point to the directory...

If you run this kernel with the new TPU Interpret Mode (by passing `interpret=pltpu.InterpretParams()` to `pl.pallas_call`), it will flag the out-of-bounds block index: ``` IndexError: Out-of-bounds block index (1, 0,...