Assorted JAX shape staging fixes for mcmc/diagnostic.py.

PiperOrigin-RevId: 398397197

tensorflow_probability/python/internal/distribution_util.py

@@ -1238,7 +1238,7 @@ def pad(x, axis, front=False, back=False, value=0, count=1, name=None):
         tensorshape_util.rank(x.shape)
         if tensorshape_util.rank(x.shape) is not None else tf.rank(
             x, name='ndims'))
-    axis = tf.convert_to_tensor(axis, name='axis')
+    axis = ps.convert_to_shape_tensor(axis, name='axis')
     axis_ = tf.get_static_value(axis)
     if axis_ is not None:
       axis = axis_

tensorflow_probability/python/mcmc/diagnostic.py

@@ -461,7 +461,7 @@ def potential_scale_reduction(chains_states,
   # array) is not efficiently computable.  Therefore, we try constant_value then
   # check for None.
   icn_const_ = tf.get_static_value(
-      tf.convert_to_tensor(independent_chain_ndims))
+      ps.convert_to_shape_tensor(independent_chain_ndims))
   if icn_const_ is not None:
     independent_chain_ndims = icn_const_
     if icn_const_ < 1:
@@ -539,15 +539,15 @@ def _potential_scale_reduction_single_state(state, independent_chain_ndims,
       state = tf.transpose(
           a=state,
           perm=ps.concat(
-              [[1, 0], tf.range(2, tf.rank(state))], axis=0))
+              [[1, 0], ps.range(2, ps.rank(state))], axis=0))
 
       # We're treating the new dim as indexing 2 chains, so increment.
       independent_chain_ndims += 1
 
-    sample_axis = tf.range(0, sample_ndims)
-    chain_axis = tf.range(sample_ndims,
+    sample_axis = ps.range(0, sample_ndims)
+    chain_axis = ps.range(sample_ndims,
                           sample_ndims + independent_chain_ndims)
-    sample_and_chain_axis = tf.range(
+    sample_and_chain_axis = ps.range(
         0, sample_ndims + independent_chain_ndims)
 
     n = _axis_size(state, sample_axis)

tensorflow_probability/python/stats/sample_stats.py

@@ -183,7 +183,7 @@ def auto_correlation(x,
     if max_lags is None:
       max_lags = x_len - 1
     else:
-      max_lags = tf.convert_to_tensor(max_lags, name='max_lags')
+      max_lags = ps.convert_to_shape_tensor(max_lags, name='max_lags')
       max_lags_ = tf.get_static_value(max_lags)
       if max_lags_ is None or not know_static_shape:
         know_static_shape = False
@@ -285,7 +285,7 @@ def cholesky_covariance(x, sample_axis=0, keepdims=False, name=None):
       lower triangular matrices (the Cholesky factors).
   """
   with tf.name_scope(name or 'cholesky_covariance'):
-    sample_axis = tf.convert_to_tensor(sample_axis, dtype=tf.int32)
+    sample_axis = ps.convert_to_shape_tensor(sample_axis, dtype=tf.int32)
     cov = covariance(
         x, sample_axis=sample_axis, event_axis=-1, keepdims=keepdims)
     return tf.linalg.cholesky(cov)
@@ -971,10 +971,10 @@ def log_average_probs(logits, sample_axis=0, event_axis=None, keepdims=False,
   with tf.name_scope(name or 'average_sigmoid'):
     logits = tf.convert_to_tensor(logits, dtype_hint=tf.float32, name='logits')
     if sample_axis is not None:
-      sample_axis = tf.convert_to_tensor(
+      sample_axis = ps.convert_to_shape_tensor(
           sample_axis, dtype_hint=tf.int32, name='sample_axis')
     if event_axis is not None:
-      event_axis = tf.convert_to_tensor(
+      event_axis = ps.convert_to_shape_tensor(
           event_axis, dtype_hint=tf.int32, name='event_axis')
     if event_axis is None:
       # log(sigmoid(x)) = log(1 / (1 + exp(-x))) = -log1p(exp(-x)) = -sp(-x)