SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1905.08205 · 2019

Towards Complex Text-to-SQL in Cross-Domain Database with Intermediate Representation

arXiv · PDF · Open in the Atlas

Code that ran

20 of 20 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 17 have no contract verdict; 3 failures of our instrument; 0 no recorded run), measured 2026-11-22; 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/1905.08205.json · badge: /api/ran/1905.08205.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
zhanzecheng/IRNet-1 canonical (classifier label)
classifier basis: census
12 ran · 0 no recorded run
brunnurs/valuenet community (classifier label)
classifier basis not recorded on the link
7 ran · 0 no recorded run
microsoft/IRNet community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
build_optimizer_encoder Ran
ran on a synthesized input
brunnurs/valuenet/src/optimizer.py
code served (permissive licence) · get_code("cfc906a590c69cbe")
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 as nn
from itertools import chain

torch.manual_seed(0)

class TransformerModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.embedding = nn.Linear(4, 8)
        self.attention = nn.Linear(8, 8)
    
    def forward(self, x):
        return self.attention(self.embedding(x))

class Encoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.transformer_model = TransformerModel()
        self.projection = nn.Linear(8, 4)
    
    def forward(self, x):
        return self.projection(self.transformer_model(x))

class DecoderCellInit(nn.Module):
    def __init__(self):
        super().__init__()
        self.init_layer = nn.Linear(4, 2)
    
    def forward(self, x):
        return self.init_layer(x)

class Model(nn.Module):
    def __init__(self):
        super().__init__()
        self.encoder = Encoder()
        self.decoder_cell_init = DecoderCellInit()
        self.output_layer = nn.Linear(2, 1)
    
    def forward(self, x):
        enc_out = self.encoder(x)
        init_out = self.decoder_cell_init(enc_out)
        return self.output_layer(init_out)

model = Model()

ARGS = {
    'model': model,
    'num_train_steps': 100,
    'lr_transformer': 1e-4,
    'lr_connection': 1e-3,
    'lr_base': 1e-2,
    'scheduler_gamma': 0.1
}
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
create_experiment_folder Ran
ran on a synthesized input
brunnurs/valuenet/src/utils.py
code served (permissive licence) · get_code("6e5bb54df87a72ba")
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 datetime
import os
import tempfile

ARGS = {
    'model_output_dir': tempfile.gettempdir(),
    'name': 'test_experiment'
}
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, 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
dot_prod_attention Ran
ran; our contract draft was wrong, not the code
zhanzecheng/IRNet-1/src/models/nn_utils.py
pointer only (licence: MIT) · get_code("d8099e5254807c28")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

dot_prod_attention Ran
ran; our contract draft was wrong, not the code
brunnurs/valuenet/src/neural_network_utils.py
code served (permissive licence) · get_code("e3de52f6de213f21")
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 torch
import torch.nn as nn

torch.manual_seed(0)

batch_size = 2
hidden_size = 4
src_sent_len = 3

h_t = torch.randn(batch_size, hidden_size)
src_encoding = torch.randn(batch_size, src_sent_len, hidden_size * 2)
src_encoding_att_linear = torch.randn(batch_size, src_sent_len, hidden_size)
mask = torch.tensor([[False, False, True], [False, True, True]], dtype=torch.uint8)

ARGS = {
    "h_t": h_t,
    "src_encoding": src_encoding,
    "src_encoding_att_linear": src_encoding_att_linear,
    "mask": mask,
}
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 — 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
dot_prod_attention Ran
ran; our contract draft was wrong, not the code
microsoft/IRNet/src/models/nn_utils.py
code served (permissive licence) · get_code("286d5b153c10befb")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Define small tensors with distinct dimensions
batch_size = 2
hidden_size = 4
src_sent_len = 3

# h_t: (batch_size, hidden_size)
h_t = torch.randn(batch_size, hidden_size)

# src_encoding: (batch_size, src_sent_len, hidden_size * 2)
src_encoding = torch.randn(batch_size, src_sent_len, hidden_size * 2)

# src_encoding_att_linear: (batch_size, src_sent_len, hidden_size)
src_encoding_att_linear = torch.randn(batch_size, src_sent_len, hidden_size)

# mask: (batch_size, src_sent_len) - optional, set to None or a valid mask
# Using None for this example, but can also be:
# mask = torch.zeros(batch_size, src_sent_len, dtype=torch.bool)
mask = None

ARGS = {
    "h_t": h_t,
    "src_encoding": src_encoding,
    "src_encoding_att_linear": src_encoding_att_linear,
    "mask": mask
}
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 — 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
fully_part_header Ran
ran on a synthesized input
zhanzecheng/IRNet-1/preprocess/utils.py
pointer only (licence: MIT) · get_code("9c3576f9e3a3fead")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_data_loader Ran
ran on a synthesized input
brunnurs/valuenet/src/data_loader.py
code served (permissive licence) · get_code("055bcd6f75fdf154")
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
from torch.utils.data import Dataset

torch.manual_seed(0)

class DummyDataset(Dataset):
    def __init__(self, size=8, num_features=4):
        self.size = size
        self.num_features = num_features
    
    def __len__(self):
        return self.size
    
    def __getitem__(self, idx):
        x = torch.randn(self.num_features)
        y = torch.tensor(idx % 2)
        return (x, y)

data_train = DummyDataset(size=8, num_features=4)
data_dev = DummyDataset(size=4, num_features=4)
batch_size = 2

ARGS = {
    "data_train": data_train,
    "data_dev": data_dev,
    "batch_size": batch_size,
    "shuffle_train": True,
    "shuffle_dev": False
}
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
get_table_colNames Ran
ran on a synthesized input
zhanzecheng/IRNet-1/src/utils.py
pointer only (licence: MIT) · get_code("963b51dd3dc75877")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

group_header Ran
ran on a synthesized input
zhanzecheng/IRNet-1/preprocess/utils.py
pointer only (licence: MIT) · get_code("9fb6cf9d7ab4b89e")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

init_config Ran
ran on a synthesized input
zhanzecheng/IRNet-1/src/args.py
pointer only (licence: MIT) · get_code("b07608731805c76b")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

length_array_to_mask_tensor Ran
ran on a synthesized input
zhanzecheng/IRNet-1/src/models/nn_utils.py
pointer only (licence: MIT) · get_code("8652ad5cc86d0e0c")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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_dataSets Ran
ran on a synthesized input
zhanzecheng/IRNet-1/preprocess/utils.py
pointer only (licence: MIT) · get_code("162d168370fb82c5")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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_word_emb Ran
ran on a synthesized input
zhanzecheng/IRNet-1/src/utils.py
pointer only (licence: MIT) · get_code("c455f0c88c1ce239")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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_word_emb Ran
ran on a synthesized input
brunnurs/valuenet/src/utils.py
code served (permissive licence) · get_code("fd6641ead78d10ee")
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 dict. 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 numpy as np
import tempfile
import os

# Create a temporary word embedding file
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "word_emb.txt")

# Write sample word embeddings to the file
with open(temp_file, 'w', encoding='utf-8') as f:
    f.write("hello 0.1 0.2 0.3 0.4\n")
    f.write("world 0.5 0.6 0.7 0.8\n")
    f.write("python 0.9 1.0 1.1 1.2\n")
    f.write("test 1.3 1.4 1.5 1.6\n")
    for i in range(100):
        f.write(f"word{i} {' '.join([str(float(j)) for j in range(4)])}\n")

ARGS = {
    'file_name': temp_file,
    'use_small': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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_word_emb_binary Ran
ran on a synthesized input
brunnurs/valuenet/src/utils.py
code served (permissive licence) · get_code("16fad9d730a01050")
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 dict. 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 codecs
import numpy as np
import tempfile
import os

# Create temporary directory and files
temp_dir = tempfile.mkdtemp()
base_path = os.path.join(temp_dir, 'test_embedding')

# Create vocab file
vocab_words = ['hello', 'world', 'test', 'embedding']
with codecs.open(base_path + '.vocab', 'w', 'utf-8') as f:
    for word in vocab_words:
        f.write(word + '\n')

# Create npy file with embeddings
embeddings = np.random.randn(len(vocab_words), 5).astype(np.float32)
np.save(base_path + '.npy', embeddings)

ARGS = {
    'embedding_file_name_w_o_suffix': base_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
lower_keys Ran
ran on a synthesized input
zhanzecheng/IRNet-1/src/utils.py
pointer only (licence: MIT) · get_code("c72d53981f7830b8")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

make_edge Ran
ran on a synthesized input
brunnurs/valuenet/src/intermediate_representation/graph.py
code served (permissive licence) · get_code("47c4fc5b3bd99c76")
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 Edge. 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
import numpy as np
from collections import deque, namedtuple

Edge = namedtuple('Edge', 'start, entry_column, end, exit_column, cost')

def make_edge(start, entry_column, end, exit_column, cost=1):
    return Edge(start, entry_column, end, exit_column, cost)

ARGS = {
    'start': 0,
    'entry_column': 1,
    'end': 2,
    'exit_column': 3,
    'cost': 5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Edge — a Edge output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Edge output the recorded ndim 1 and dtype family int describe that coercion, not the output, and are not 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
pop_front Ran
ran on a synthesized input
zhanzecheng/IRNet-1/sem2SQL.py
pointer only (licence: MIT) · get_code("556a6efd5f06ab91")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

split_logical_form Ran
ran on a synthesized input
zhanzecheng/IRNet-1/sem2SQL.py
pointer only (licence: MIT) · get_code("beb755eade641481")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

table_dict_to_mask_tensor Ran
ran on a synthesized input
zhanzecheng/IRNet-1/src/models/nn_utils.py
pointer only (licence: MIT) · get_code("7e01be511a9ebcf6")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, zhanzecheng/IRNet-1, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 19 distinct functions with an identity, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 2 later) and 17 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
dot_prod_attention 2 other papers
1 dated earlier, 1 later
1810.02720 (2018-10)
TRANX: A Transition-based Neural Abstract Syntax Parser for Semantic Parsing and Code Generation
2004.09015 (2020-04)
Incorporating External Knowledge through Pre-training for Natural Language to Code Generation
dot_prod_attention 1 other paper
1 later
2105.14796 (2021-05)
Improving Tree-Structured Decoder Training for Code Generation via Mutual Learning

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 2026-11-22.

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 present a neural approach called IRNet for complex and cross-domain Text-to-SQL. IRNet aims to address two challenges: 1) the mismatch between intents expressed in natural language (NL) and the implementation details in SQL; 2) the challenge in predicting columns caused by the large number of out-of-domain words. Instead of end-to-end synthesizing a SQL query, IRNet decomposes the synthesis process into three phases. In the first phase, IRNet performs a schema linking over a question and a database schema. Then, IRNet adopts a grammar-based neural model to synthesize a SemQL query which is an intermediate representation that we design to bridge NL and SQL. Finally, IRNet deterministically infers a SQL query from the synthesized SemQL query with domain knowledge. On the challenging Text-to-SQL benchmark Spider, IRNet achieves 46.7% accuracy, obtaining 19.5% absolute improvement over previous state-of-the-art approaches. At the time of writing, IRNet achieves the first position on the Spider leaderboard.

For agents

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

get_harvested_code_for_paper("1905.08205")
get_code_for_paper("1905.08205")
have("1905.08205")

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

curl https://syntology.ai/api/ran/1905.08205.json

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

[![Syntology run record](https://syntology.ai/api/ran/1905.08205.svg)](https://syntology.ai/paper/1905.08205)

Connect an agent — have() is free.