Trivial modifications

zysite · zysite · commit 845bf056f976 · 2019-12-11T04:40:15.000Z
diff --git a/parser/modules/__init__.py b/parser/modules/__init__.py
@@ -7,5 +7,5 @@
 from .char_lstm import CHAR_LSTM
 from .mlp import MLP
 
-__all__ = ['MLP', 'BertEmbedding', 'Biaffine',
-           'BiLSTM', 'CHAR_LSTM', 'dropout']
+__all__ = ['CHAR_LSTM', 'MLP', 'BertEmbedding',
+           'Biaffine', 'BiLSTM', 'dropout']
diff --git a/parser/modules/bilstm.py b/parser/modules/bilstm.py
@@ -81,8 +81,7 @@ def layer_forward(self, x, hx, cell, batch_sizes, reverse=False):
             output.reverse()
         else:
             hx_n.append(hx_i)
-            hx_n.reverse()
-            hx_n = [torch.cat(h) for h in zip(*hx_n)]
+            hx_n = [torch.cat(h) for h in zip(*reversed(hx_n))]
         output = torch.cat(output)
 
         return output, hx_n
diff --git a/parser/utils/alg.py b/parser/utils/alg.py
@@ -15,9 +15,7 @@ def kmeans(x, k):
     # assign labels to each datapoint based on centroids
     dists, y = torch.abs_(d.unsqueeze(-1) - c).min(dim=-1)
     # make sure number of datapoints is greater than that of clusters
-    if len(d) < k:
-        raise AssertionError(f"unable to assign {len(d)} datapoints to "
-                             f"{k} clusters")
+    assert len(d) >= k, f"unable to assign {len(d)} datapoints to {k} clusters"
 
     while old is None or not c.equal(old):
         # if an empty cluster is encountered,
@@ -59,27 +57,27 @@ def eisner(scores, mask):
     for w in range(1, seq_len):
         n = seq_len - w
         starts = p_i.new_tensor(range(n)).unsqueeze(0)
-        # ilr = C(i, r) + C(j, r+1)
+        # ilr = C(i->r) + C(j->r+1)
         ilr = stripe(s_c, n, w) + stripe(s_c, n, w, (w, 1))
         # [batch_size, n, w]
         ilr = ilr.permute(2, 0, 1)
         il = ilr + scores.diagonal(-w).unsqueeze(-1)
-        # I(j, i) = max(C(i, r) + C(j, r+1) + S(j, i)), i <= r < j
+        # I(j->i) = max(C(i->r) + C(j->r+1) + s(j->i)), i <= r < j
         il_span, il_path = il.max(-1)
         s_i.diagonal(-w).copy_(il_span)
         p_i.diagonal(-w).copy_(il_path + starts)
         ir = ilr + scores.diagonal(w).unsqueeze(-1)
-        # I(i, j) = max(C(i, r) + C(j, r+1) + S(i, j)), i <= r < j
+        # I(i->j) = max(C(i->r) + C(j->r+1) + s(i->j)), i <= r < j
         ir_span, ir_path = ir.max(-1)
         s_i.diagonal(w).copy_(ir_span)
         p_i.diagonal(w).copy_(ir_path + starts)
 
-        # C(j, i) = max(C(r, i) + I(j, r)), i <= r < j
-        cl = stripe(s_c, n, w, dim=0) + stripe(s_i, n, w, (w, 0))
+        # C(j->i) = max(C(r->i) + I(j->r)), i <= r < j
+        cl = stripe(s_c, n, w, (0, 0), 0) + stripe(s_i, n, w, (w, 0))
         cl_span, cl_path = cl.permute(2, 0, 1).max(-1)
         s_c.diagonal(-w).copy_(cl_span)
         p_c.diagonal(-w).copy_(cl_path + starts)
-        # C(i, j) = max(I(i, r) + C(r, j)), i < r <= j
+        # C(i->j) = max(I(i->r) + C(r->j)), i < r <= j
         cr = stripe(s_i, n, w, (0, 1)) + stripe(s_c, n, w, (1, w), 0)
         cr_span, cr_path = cr.permute(2, 0, 1).max(-1)
         s_c.diagonal(w).copy_(cr_span)
@@ -136,7 +134,7 @@ def stripe(x, n, w, offset=(0, 0), dim=1):
     tensor([[ 0,  5, 10],
             [ 6, 11, 16]])
     '''
-    seq_len = x.size(1)
+    x, seq_len = x.contiguous(), x.size(1)
     stride, numel = list(x.stride()), x[0, 0].numel()
     stride[0] = (seq_len + 1) * numel
     stride[1] = (1 if dim == 1 else seq_len) * numel
diff --git a/parser/utils/corpus.py b/parser/utils/corpus.py
@@ -76,7 +76,7 @@ def load(cls, path, fields):
             lines = [line.strip() for line in f]
         for i, line in enumerate(lines):
             if not line:
-                values = list(zip(*[l.split() for l in lines[start:i]]))
+                values = list(zip(*[l.split('\t') for l in lines[start:i]]))
                 sentences.append(Sentence(fields, values))
                 start = i + 1
 
diff --git a/parser/utils/metric.py b/parser/utils/metric.py
@@ -12,7 +12,7 @@ def __init__(self, eps=1e-5):
         self.correct_rels = 0.0
 
     def __repr__(self):
-        return f"UAS: {self.uas:.2%} LAS: {self.las:.2%}"
+        return f"UAS: {self.uas:6.2%} LAS: {self.las:6.2%}"
 
     def __call__(self, arc_preds, rel_preds, arc_golds, rel_golds, mask):
         arc_mask = arc_preds.eq(arc_golds)[mask]
