Home › Census › MultiHeadAttention

MultiHeadAttention

Syntologyfunction-name censuscensus 2026-09-22battery b986f7e04d79all samples with this name

MultiHeadAttention: 9 implementations from 7 papers ran on one shared input (census 2026-09-22, battery b986f7e04d79); they produced 9 distinct outputs across 3 buckets, one shared input per bucket.

Identical values to six decimals (the recorded digest) on the shared input are agreement on those inputs, not equivalence. Implementations are compared only within one bucket, the positional (rank, kind, dtype) of each array argument; the argument name is not part of the key because the harness draws the shared array from (rank, kind) and casts it to the dtype, whatever the name; each bucket's shared input is fixed by that key, so members of one bucket saw bitwise-identical inputs under their own scalar arguments. A cluster is the set of members whose recorded output digest is identical. Nothing here says which computation a paper's method intended, and nothing reproduces a paper's results.

Not compared, and not in the tables or the counts above:

Bucket 1 of 3: arg 1: rank 3, kind float, dtype float32 · arg 2: rank 3, kind float, dtype float32 · arg 3: rank 3, kind float, dtype float32

4 implementations from 4 papers share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 4 distinct outputs, largest cluster first. Values are the first 8 of the recorded output, flattened.

Cluster (same digest to six decimals)MembersShared output on this bucket's input
1 implementation
1 paper
0cabec910b03
class
libfewshot_core/model/metric/feat.py abb0132a
recorded q:3/float/float32, k:3/float/float32, v:3/float/float32
[-1.26393, 1.92103, -0.729993, 0.608796, 0.419665, -0.96795, 0.595852, -0.583477, …]
shape [2, 4, 8] · float32 · Tensor
1 implementation
1 paper
58fd5c1158b2
class
src/nlp/layers/encoder.py 9c478f15
recorded q:3/float/float32, k:3/float/float32, v:3/float/float32; mask=None
[-0, 1.32319, -1.21911, -0, 0.189581, -0.886993, -0.304486, 0, …]
shape [2, 4, 8] · float32 · Tensor
1 implementation
1 paper
b47c72b25d74
class
summarizer_transformer.py db4e1c41
recorded q:3/float/float32, k:3/float/float32, v:3/float/float32; mask=False, cover=None
[0.035665, -0.413472, -0.177767, -0.208916, 0.23824, 0.435368, 0.0700511, 0.0641225, …]
shape [2, 4, 8] · float32 · Tensor
1 implementation
1 paper
f1e879c2058d
class
bpp/model.py 337e5971
recorded query:3/float/float32, key:3/float/float32, value:3/float/float32; mask=None
[0.240397, -0.0802591, -0.205289, 0.0541918, -0.292073, 0.605929, -0.417172, -0.00929932, …]
shape [2, 4, 8] · float32 · Tensor

Bucket 2 of 3: arg 1: rank 3, kind float, dtype float32

4 implementations from 2 papers share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 4 distinct outputs, largest cluster first. Values are the first 8 of the recorded output, flattened.

Cluster (same digest to six decimals)MembersShared output on this bucket's input
1 implementation
1 paper
4dfb8cb5987d
class
models/mae.py f6a1ed2d
recorded x:3/float/float32
[0.534887, 0.247349, 0.471544, -0.00193934, -0.292215, -0.146134, 0.290468, -0.0427631, …]
shape [2, 4, 8] · float32 · Tensor
1 implementation
1 paper
57f8a2a43047
class
mobileclip/clip.py 88f8bcf4
recorded x_q:3/float/float32; x_kv=None, key_padding_mask=None, attn_mask=None
[-0.0803661, -0.117569, -0.0700971, 0.24619, 0.236701, 0.305764, -0.00580824, 0.411263, …]
shape [2, 4, 8] · float32 · Tensor
1 implementation
1 paper
a8d86eb7e123
class
vision_transformer.py 6082f915
recorded x:3/float/float32
[-0.171513, 0.542943, -0.0983858, -0.315013, 0.1322, 0.576283, -0.350209, 1.13742, …]
shape [2, 4, 8] · float32 · Tensor
1 implementation
1 paper
c60576142d8a
class
src/ViT.py 6bad9911
recorded x:3/float/float32
[0.18117, -0.0366058, -0.364896, 0.00403607, -0.253827, 0.626424, -0.379873, -0.0199547, …]
shape [2, 4, 8] · float32 · Tensor

Bucket 3 of 3: arg 1: rank 3, kind float, dtype float32 · arg 2: rank 3, kind float, dtype float32

1 implementation from 1 paper share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 1 distinct output, largest cluster first. Values are the first 8 of the recorded output, flattened.

Cluster (same digest to six decimals)MembersShared output on this bucket's input
1 implementation
1 paper
8132fe143af3
class
nets/attention_model.py 3e9c7f43
recorded q:3/float/float32, h:3/float/float32; mask=None
[0.149484, 0.172254, -0.103393, 0.0170991, -0.14438, -0.105019, 0.102146, -0.14146, …]
shape [2, 4, 16] · float32 · Tensor

Identical values to six decimals (the recorded digest) on the shared input are agreement on those inputs, not equivalence; where a cluster's members carry recorded values, the largest difference among them is shown under the cluster. Paper titles are the archive's archive 2025-07-28 where the paper is in the archive and the graph's where it was added by Syntology; papers with no page here are shown by their recorded paper id only. A paper count above the implementation count means one implementation (one code sha) is held from several papers' repositories and counts once. Per-sample status, licence and fingerprint records for each paper are on its paper page. JSON twin: /census/multiheadattention.json.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections