Update diagnostics.py (#254)

* update diagnostics.py

* do some changes

icefall/diagnostics.py

@@ -1,5 +1,6 @@
 # Copyright      2022  Xiaomi Corp.        (authors: Daniel Povey
-#                                                    Zengwei Yao)
+#                                                    Zengwei Yao
+#                                                    Mingshuang Luo)
 #
 # See ../LICENSE for clarification regarding multiple authors
 #
@@ -17,7 +18,7 @@
 
 
 import random
-from typing import List, Tuple
+from typing import List, Optional, Tuple
 
 import torch
 from torch import Tensor, nn
@@ -28,22 +29,29 @@ class TensorDiagnosticOptions(object):
 
     Args:
       memory_limit:
-        The maximum number of bytes per tensor (limits how many copies
+        The maximum number of bytes per tensor
-        of the tensor we cache).
+        (limits how many copies of the tensor we cache).
+      max_eig_dim:
+        The maximum dimension for which we print out eigenvalues
+        (limited for speed reasons).
     """
 
-    def __init__(self, memory_limit: int):
+    def __init__(self, memory_limit: int = (2 ** 20), max_eig_dim: int = 512):
         self.memory_limit = memory_limit
+        self.max_eig_dim = max_eig_dim
 
     def dim_is_summarized(self, size: int):
         return size > 10 and size != 31
 
 
-def get_sum_abs_stats(
+def get_tensor_stats(
-    x: Tensor, dim: int, stats_type: str
+    x: Tensor,
+    dim: int,
+    stats_type: str,
 ) -> Tuple[Tensor, int]:
-    """Returns the sum-of-absolute-value of this Tensor, for each index into
+    """
-    the specified axis/dim of the tensor.
+    Returns the specified transformation of the Tensor (either x or x.abs()
+    or (x > 0), summed over all but the index `dim`.
 
     Args:
       x:
@@ -51,28 +59,38 @@ def get_sum_abs_stats(
       dim:
         Dimension with 0 <= dim < x.ndim
       stats_type:
-        Either "mean-abs" in which case the stats represent the mean absolute
+        The stats_type includes several types:
-        value, or "pos-ratio" in which case the stats represent the proportion
+        "abs" -> take abs() before summing
-        of positive values (actually: the tensor is count of positive values,
+        "positive" -> take (x > 0) before summing
-        count is the count of all values).
+        "rms" -> square before summing, we'll take sqrt later
-
+        "value -> just sum x itself
     Returns:
-      (sum_abs, count) where sum_abs is a Tensor of shape (x.shape[dim],),
+      stats: a Tensor of shape (x.shape[dim],).
-      and the count is an integer saying how many items were counted in
+      count: an integer saying how many items were counted in each element
-      each element of sum_abs.
+      of stats.
     """
-    if stats_type == "mean-abs":
+
+    count = x.numel() // x.shape[dim]
+
+    if stats_type == "eigs":
+        x = x.transpose(dim, -1)
+        x = x.reshape(-1, x.shape[-1])
+        # shape of returned tensor: (s, s),
+        # where s is size of dimension `dim` of original x.
+        return torch.matmul(x.transpose(0, 1), x), count
+    elif stats_type == "abs":
         x = x.abs()
-    else:
+    elif stats_type == "rms":
-        assert stats_type == "pos-ratio"
+        x = x ** 2
+    elif stats_type == "positive":
         x = (x > 0).to(dtype=torch.float)
+    else:
+        assert stats_type == "value"
 
-    orig_numel = x.numel()
     sum_dims = [d for d in range(x.ndim) if d != dim]
+    if len(sum_dims) > 0:
         x = torch.sum(x, dim=sum_dims)
-    count = orig_numel // x.numel()
     x = x.flatten()
-
     return x, count
 
 
@@ -83,43 +101,58 @@ def get_diagnostics_for_dim(
     sizes_same: bool,
     stats_type: str,
 ) -> str:
-    """This function gets diagnostics for a dimension of a module.
+    """
+    This function gets diagnostics for a dimension of a module.
 
     Args:
       dim:
-        The dimension to analyze, with 0 <= dim < tensors[0].ndim
+        the dimension to analyze, with 0 <= dim < tensors[0].ndim
-      tensors:
-        List of cached tensors to get the stats
       options:
-        Options object
+        options object
       sizes_same:
         True if all the tensor sizes are the same on this dimension
-        stats_type: either "mean-abs" or "pos-ratio", dictates the type of
+        stats_type: either "abs" or "positive" or "eigs" or "value",
-        stats we accumulate, mean-abs is mean absolute value, "pos-ratio" is
+        imdictates the type of stats we accumulate, abs is mean absolute
-        proportion of positive to nonnegative values.
+        value, "positive" is proportion of positive to nonnegative values,
-
+        "eigs" is eigenvalues after doing outer product on this dim, sum
+        over all other dimes.
     Returns:
       Diagnostic as a string, either percentiles or the actual values,
-      see the code.
+      see the code.  Will return the empty string if the diagnostics did
+      not make sense to print out for this dimension, e.g. dimension
+      mismatch and stats_type == "eigs".
     """
 
     # stats_and_counts is a list of pair (Tensor, int)
-    stats_and_counts = [get_sum_abs_stats(x, dim, stats_type) for x in tensors]
+    stats_and_counts = [get_tensor_stats(x, dim, stats_type) for x in tensors]
     stats = [x[0] for x in stats_and_counts]
     counts = [x[1] for x in stats_and_counts]
-    if sizes_same:
+
+    if stats_type == "eigs":
+        try:
+            stats = torch.stack(stats).sum(dim=0)
+        except:  # noqa
+            return ""
+        count = sum(counts)
+        stats = stats / count
+        stats, _ = torch.symeig(stats)
+        stats = stats.abs().sqrt()
+        # sqrt so it reflects data magnitude, like stddev- not variance
+    elif sizes_same:
         stats = torch.stack(stats).sum(dim=0)
         count = sum(counts)
         stats = stats / count
     else:
         stats = [x[0] / x[1] for x in stats_and_counts]
         stats = torch.cat(stats, dim=0)
+    if stats_type == "rms":
+        stats = stats.sqrt()
 
-    # If `summarize` we print percentiles of the stats;
+    # if `summarize` we print percentiles of the stats; else,
-    # else, we print out individual elements.
+    # we print out individual elements.
     summarize = (not sizes_same) or options.dim_is_summarized(stats.numel())
     if summarize:
-        # Print out percentiles.
+        # print out percentiles.
         stats = stats.sort()[0]
         num_percentiles = 10
         size = stats.numel()
@@ -129,12 +162,25 @@ def get_diagnostics_for_dim(
             percentiles.append(stats[index].item())
         percentiles = ["%.2g" % x for x in percentiles]
         percentiles = " ".join(percentiles)
-        return f"percentiles: [{percentiles}]"
+        ans = f"percentiles: [{percentiles}]"
     else:
-        stats = stats.tolist()
+        ans = stats.tolist()
-        stats = ["%.2g" % x for x in stats]
+        ans = ["%.2g" % x for x in ans]
-        stats = "[" + " ".join(stats) + "]"
+        ans = "[" + " ".join(ans) + "]"
-        return stats
+    if stats_type == "value":
+        # This norm is useful because it is strictly less than the largest
+        # sqrt(eigenvalue) of the variance, which we print out, and shows,
+        # speaking in an approximate way, how much of that largest eigenvalue
+        # can be attributed to the mean of the distribution.
+        norm = (stats ** 2).sum().sqrt().item()
+        mean = stats.mean().item()
+        rms = (stats ** 2).mean().sqrt().item()
+        ans += f", norm={norm:.2g}, mean={mean:.2g}, rms={rms:.2g}"
+    else:
+        mean = stats.mean().item()
+        rms = (stats ** 2).mean().sqrt().item()
+        ans += f", mean={mean:.2g}, rms={rms:.2g}"
+    return ans
 
 
 def print_diagnostics_for_dim(
@@ -153,17 +199,27 @@ def print_diagnostics_for_dim(
         Options object.
     """
 
-    for stats_type in ["mean-abs", "pos-ratio"]:
+    ndim = tensors[0].ndim
-        # stats_type will be "mean-abs" or "pos-ratio".
+    if ndim > 1:
+        stats_types = ["abs", "positive", "value", "rms"]
+        if tensors[0].shape[dim] <= options.max_eig_dim:
+            stats_types.append("eigs")
+    else:
+        stats_types = ["value", "abs"]
+
+    for stats_type in stats_types:
         sizes = [x.shape[dim] for x in tensors]
         sizes_same = all([x == sizes[0] for x in sizes])
         s = get_diagnostics_for_dim(
             dim, tensors, options, sizes_same, stats_type
         )
+        if s == "":
+            continue
 
         min_size = min(sizes)
         max_size = max(sizes)
         size_str = f"{min_size}" if sizes_same else f"{min_size}..{max_size}"
+        # stats_type will be "abs" or "positive".
         print(f"module={name}, dim={dim}, size={size_str}, {stats_type} {s}")
 
 
@@ -225,11 +281,15 @@ class TensorDiagnostic(object):
             # Ensure there is at least one dim.
             self.saved_tensors = [x.unsqueeze(0) for x in self.saved_tensors]
 
+        try:
+            device = torch.device("cuda")
+        except:  # noqa
+            device = torch.device("cpu")
+
         ndim = self.saved_tensors[0].ndim
+        tensors = [x.to(device) for x in self.saved_tensors]
         for dim in range(ndim):
-            print_diagnostics_for_dim(
+            print_diagnostics_for_dim(self.name, dim, tensors, self.opts)
-                self.name, dim, self.saved_tensors, self.opts
-            )
 
 
 class ModelDiagnostic(object):
@@ -240,11 +300,14 @@ class ModelDiagnostic(object):
         Options object.
     """
 
-    def __init__(self, opts: TensorDiagnosticOptions):
+    def __init__(self, opts: Optional[TensorDiagnosticOptions] = None):
         # In this dictionary, the keys are tensors names and the values
         # are corresponding TensorDiagnostic objects.
-        self.diagnostics = dict()
+        if opts is None:
+            self.opts = TensorDiagnosticOptions()
+        else:
             self.opts = opts
+        self.diagnostics = dict()
 
     def __getitem__(self, name: str):
         if name not in self.diagnostics:
@@ -321,7 +384,7 @@ def attach_diagnostics(
 
 
 def _test_tensor_diagnostic():
-    opts = TensorDiagnosticOptions(2 ** 20)
+    opts = TensorDiagnosticOptions(2 ** 20, 512)
 
     diagnostic = TensorDiagnostic(opts, "foo")