SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2008.07772 · 2020

Very Deep Transformers for Neural Machine Translation

arXiv · PDF · Open in the Atlas

Code that ran

8 of 9 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 6 failures of our instrument; 1 no recorded run), measured 2026-11-21; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2008.07772.json · badge: /api/ran/2008.07772.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
microsoft/deepnmt community (classifier label)
classifier basis: plan_row
5 ran · 1 no recorded run
LiyuanLucasLiu/Transformer-Clinic extension (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
Embedding Ran
ran; our contract draft was wrong, not the code · constructed an object; no result was computed
microsoft/deepnmt/deepnmt/adv_transformer.py
pointer only (licence: MIT); the stored copy get_code serves, from instance-wise-ordered-transformer/iot, is pointer only (licence: NOASSERTION) · get_code("96cbb5e9ca5b6be0")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from instance-wise-ordered-transformer/iot, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

KL Ran
ran on a synthesized input
microsoft/deepnmt/deepnmt/adv_label_smoothed_cross_entropy.py
code served (permissive licence) · get_code("c8d783d775fd71be")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for input and target
# input should be log-probabilities (output of log_softmax)
# target can be logits or probabilities
batch_size = 2
num_classes = 4

input_tensor = torch.randn(batch_size, num_classes)
target_tensor = torch.randn(batch_size, num_classes)

ARGS = {
    'input': input_tensor,
    'target': target_tensor,
    'reduction': 'sum'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
Linear Ran
ran; our contract draft was wrong, not the code · constructed an object; no result was computed
microsoft/deepnmt/deepnmt/transformer_layer.py
pointer only (licence: MIT); the stored copy get_code serves, from instance-wise-ordered-transformer/iot, is pointer only (licence: NOASSERTION) · get_code("8cd8cb0d1e9e63e4")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from instance-wise-ordered-transformer/iot, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

dictolist Ran
ran; our contract draft was wrong, not the code
LiyuanLucasLiu/Transformer-Clinic/fairseq/examples/translation_moe/score.py
code served (permissive licence) · get_code("81e9530973deb9b1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

d = {
    'zebra': 42,
    'apple': 'hello',
    'monkey': [1, 2, 3],
    'banana': {'nested': 'dict'}
}

ARGS = {
    'd': d
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
label_smoothed_nll_loss Ran
ran on a synthesized input
microsoft/deepnmt/deepnmt/adv_label_smoothed_cross_entropy.py
code served (permissive licence) · get_code("0084024f3f3086ec")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# Construct lprobs: log probabilities over vocabulary
# Shape: (batch_size=2, seq_len=3, vocab_size=4)
# These should be log probabilities (negative values from log_softmax)
lprobs = torch.randn(2, 3, 4)
lprobs = F.log_softmax(lprobs, dim=-1)

# Construct target: token indices
# Shape: (batch_size=2, seq_len=3)
# Values should be in range [0, vocab_size)
target = torch.randint(0, 4, (2, 3))

# epsilon: label smoothing factor (typically small, e.g., 0.1)
epsilon = 0.1

# ignore_index: optional index to ignore (e.g., padding token)
# Using None to test the branch without masking first
ignore_index = None

# reduce: whether to reduce loss to scalar
reduce = True

ARGS = {
    "lprobs": lprobs,
    "target": target,
    "epsilon": epsilon,
    "ignore_index": ignore_index,
    "reduce": reduce,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
last_n_checkpoints Ran
ran; our contract draft was wrong, not the code
microsoft/deepnmt/nmt_eval/average_checkpoints.py
pointer only (licence: MIT); the stored copy get_code serves, from ubc-nlp/llmas, is pointer only (licence: NOASSERTION) · get_code("99de76104bad1fc9")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ubc-nlp/llmas, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

load_ref Ran
ran; our contract draft was wrong, not the code
LiyuanLucasLiu/Transformer-Clinic/fairseq/examples/translation_moe/score.py
code served (permissive licence) · get_code("3ed3082e1ace0b17")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import os

# Create a temporary file with the expected format
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_path = temp_file.name

# Write test data in the expected format
content = """S-1	source sentence one
T-1	target sentence one
R-1	reference one a
R-2	reference one b
S-2	source sentence two
T-2	target sentence two
R-1	reference two a
"""

temp_file.write(content)
temp_file.close()

ARGS = {
    'path': temp_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_sys Ran
ran; our contract draft was wrong, not the code
LiyuanLucasLiu/Transformer-Clinic/fairseq/examples/translation_moe/score.py
code served (permissive licence) · get_code("874a757bff402af6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import os

# Create temporary files with the expected format
temp_dir = tempfile.mkdtemp()
test_file = os.path.join(temp_dir, "test.txt")

# Write test data in the expected format
with open(test_file, 'w') as f:
    # Source sentences (S-)
    f.write("S-0\tHello world\n")
    f.write("S-1\tGoodbye world\n")
    # Target sentences (T-)
    f.write("T-0\tBonjour monde\n")
    f.write("T-1\tAu revoir monde\n")
    # Hypotheses (H-) with log probabilities
    f.write("H-0\t-1.5\tHello world\n")
    f.write("H-0\t-2.3\tHi world\n")
    f.write("H-1\t-0.8\tGoodbye world\n")
    f.write("H-1\t-1.2\tBye world\n")

ARGS = {
    "paths": [test_file]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
average_checkpoints Not yet run
unverified (harvested, no recorded run)
microsoft/deepnmt/nmt_eval/average_checkpoints.py
code served (permissive licence) · get_code("7cc2e301e30389d1")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 4 later) and 5 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
Linear 58 other papers
11 dated earlier, 47 later · 5 only as an assembly brick
1409.0473 (2014-09)
Neural Machine Translation by Jointly Learning to Align and Translate
1706.03762 (2017-06)
Attention Is All You Need
1904.01681 (2019-04)
Augmented Neural ODEs
and 55 more not named here
Embedding 29 other papers
5 dated earlier, 24 later
1904.01681 (2019-04)
Augmented Neural ODEs
1906.12183 (2019-06)
Neural ODEs as the Deep Limit of ResNets with constant weights
2002.03184 (2020-02)
Time-aware Large Kernel Convolutions
and 26 more not named here
average_checkpoints 4 other papers
1 dated earlier, 3 later
2002.03184 (2020-02)
Time-aware Large Kernel Convolutions
2210.07535 (2022-10)
AutoMoE: Heterogeneous Mixture-of-Experts with Adaptive Computation for Efficient Neural Machine Translation
2310.16712 (2023-10)
LLM Performance Predictors are good initializers for Architecture Search
and 1 more not named here
last_n_checkpoints 3 other papers
1 dated earlier, 2 later
2002.03184 (2020-02)
Time-aware Large Kernel Convolutions
2210.07535 (2022-10)
AutoMoE: Heterogeneous Mixture-of-Experts with Adaptive Computation for Efficient Neural Machine Translation
2310.16712 (2023-10)
LLM Performance Predictors are good initializers for Architecture Search

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We explore the application of very deep Transformer models for Neural Machine Translation (NMT). Using a simple yet effective initialization technique that stabilizes training, we show that it is feasible to build standard Transformer-based models with up to 60 encoder layers and 12 decoder layers. These deep models outperform their baseline 6-layer counterparts by as much as 2.5 BLEU, and achieve new state-of-the-art benchmark results on WMT14 English-French (43.8 BLEU and 46.4 BLEU with back-translation) and WMT14 English-German (30.1 BLEU).The code and trained models will be publicly available at: https://github.com/namisan/exdeep-nmt.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2008.07772")
get_code_for_paper("2008.07772")
have("2008.07772")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2008.07772.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2008.07772.svg)](https://syntology.ai/paper/2008.07772)

Connect an agent — have() is free.