SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.05095 · 2021

ST++: Make Self-training Work Better for Semi-supervised Semantic Segmentation

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 (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 5 failures of our instrument; 1 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/2106.05095.json · badge: /api/ran/2106.05095.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
LiheYoung/ST-PlusPlus canonical (classifier label)
classifier basis: census
8 ran · 1 no recorded run
FunctionStatusWhere it lives
ASPPConv Ran
ran; our contract draft was wrong, not the code
LiheYoung/ST-PlusPlus/model/semseg/deeplabv3plus.py
pointer only (licence: MIT); the stored copy get_code serves, from cvlab-stonybrook/beyond-pixels, is pointer only (licence: NONE) · get_code("392d0834780485a1")
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 cvlab-stonybrook/beyond-pixels, and GitHub detected no licence in that repository when it was checked. 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.

color_map Ran
ran on a synthesized input
LiheYoung/ST-PlusPlus/utils.py
code served (permissive licence) · get_code("07d14767c4ecb04a")
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 ndarray, ndim 2, dtype family int, shape (256, 3) 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 numpy as np

ARGS = {
    'dataset': 'pascal'
}
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 ndarray, ndim 2, dtype family int, shape (256, 3) — shape (256, 3) 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
conv1x1 Ran
ran; our contract draft was wrong, not the code
LiheYoung/ST-PlusPlus/model/backbone/resnet.py
pointer only (licence: MIT); the stored copy get_code serves, from guanyuezhen/ufcd, is pointer only (licence: NONE) · get_code("6ef392f365cc5703")
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 guanyuezhen/ufcd, and GitHub detected no licence in that repository when it was checked. 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.

conv3x3 Ran
ran; our contract draft was wrong, not the code
LiheYoung/ST-PlusPlus/model/backbone/resnet.py
pointer only (licence: MIT); the stored copy get_code serves, from lewandofskee/MambaAD, is pointer only (licence: NONE) · get_code("794a2ed91a5175bd")
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 lewandofskee/MambaAD, and GitHub detected no licence in that repository when it was checked. 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.

count_params Ran
ran, honoured the contract we drafted
LiheYoung/ST-PlusPlus/utils.py
code served (permissive licence) · get_code("c9896e2a9e9768ef")
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 float, 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 as nn

torch.manual_seed(0)

# Create a minimal model with some parameters
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    "model": model
}
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 float, 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
crop Ran
ran on a synthesized input
LiheYoung/ST-PlusPlus/dataset/transform.py
code served (permissive licence) · get_code("52fb4f1210c91ce6")
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.

from PIL import Image
import random

random.seed(0)

# Create a small test image (RGB mode)
img = Image.new('RGB', (8, 6), color=(100, 150, 200))

# Create a small test mask (L mode for grayscale)
mask = Image.new('L', (8, 6), color=(128,))

# Target crop size
size = 5

ARGS = {
    'img': img,
    'mask': mask,
    'size': size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — 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
hflip Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
LiheYoung/ST-PlusPlus/dataset/transform.py
code served (permissive licence) · get_code("49f7eaf0ab330e0d")
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.

from PIL import Image
import random

# Create small PIL Image objects for img and mask
# The function expects PIL Image objects and calls .transpose() on them
img = Image.new('RGB', (4, 3), color='red')
mask = Image.new('L', (4, 3), color=128)

# Set random seed for reproducibility
random.seed(0)

ARGS = {
    'img': img,
    'mask': mask,
    'p': 0.5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — 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
normalize Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
LiheYoung/ST-PlusPlus/dataset/transform.py
code served (permissive licence) · get_code("4d761942e03d5f8b")
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 numpy as np
import torch
from torchvision import transforms
from PIL import Image

torch.manual_seed(0)

# Create a small PIL image (4x3 RGB)
img_array = np.random.randint(0, 256, (4, 3, 3), dtype=np.uint8)
img = Image.fromarray(img_array, mode='RGB')

# Create a small PIL mask (4x3 grayscale)
mask_array = np.random.randint(0, 10, (4, 3), dtype=np.uint8)
mask = Image.fromarray(mask_array, mode='L')

ARGS = {
    "img": img,
    "mask": mask
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, torch, torchvision — 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
resnet18 Not yet run
unverified (harvested, no recorded run)
LiheYoung/ST-PlusPlus/model/backbone/resnet.py
code served (permissive licence) · get_code("64a129d9f760a0da")
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, 7 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 2 the same month, 7 later) and 2 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
conv1x1 431 other papers
132 dated earlier, 5 the same month, 288 later, 5 same year, order not determinable, 1 undated · 38 only as an assembly brick
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1506.01186 (2015-06)
Cyclical Learning Rates for Training Neural Networks
1507.05717 (2015-07)
An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition
and 428 more not named here
conv3x3 312 other papers
80 dated earlier, 2 the same month, 226 later, 3 same year, order not determinable, 1 undated · 31 only as an assembly brick
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1507.05717 (2015-07)
An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
and 309 more not named here
hflip 6 other papers
6 later
2208.09910 (2022-08)
Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation
2308.09281 (2023-08)
Diverse Cotraining Makes Strong Semi-Supervised Segmentor
2403.06462 (2024-03)
Towards the Uncharted: Density-Descending Feature Perturbation for Semi-supervised Semantic Segmentation
and 3 more not named here
normalize 5 other papers
5 later
2208.09910 (2022-08)
Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation
2308.09281 (2023-08)
Diverse Cotraining Makes Strong Semi-Supervised Segmentor
2405.20610 (2024-05)
PrevMatch: Revisiting and Maximizing Temporal Knowledge in Semi-Supervised Semantic Segmentation
and 2 more not named here
ASPPConv 3 other papers
3 later
2208.09910 (2022-08)
Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation
2405.20610 (2024-05)
PrevMatch: Revisiting and Maximizing Temporal Knowledge in Semi-Supervised Semantic Segmentation
2407.04036 (2024-07)
Beyond Pixels: Semi-Supervised Semantic Segmentation with a Multi-scale Patch-based Multi-Label Classifier
count_params 2 other papers
2 later
2401.08407 (2024-01)
Cross-Domain Few-Shot Segmentation via Iterative Support-Query Correspondence Mining
2404.08951 (2024-04)
Constructing and Exploring Intermediate Domains in Mixed Domain Semi-supervised Medical Image Segmentation
resnet18 1 other paper
1 later
2401.08407 (2024-01)
Cross-Domain Few-Shot Segmentation via Iterative Support-Query Correspondence Mining

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

Self-training via pseudo labeling is a conventional, simple, and popular pipeline to leverage unlabeled data. In this work, we first construct a strong baseline of self-training (namely ST) for semi-supervised semantic segmentation via injecting strong data augmentations (SDA) on unlabeled images to alleviate overfitting noisy labels as well as decouple similar predictions between the teacher and student. With this simple mechanism, our ST outperforms all existing methods without any bells and whistles, e.g., iterative re-training. Inspired by the impressive results, we thoroughly investigate the SDA and provide some empirical analysis. Nevertheless, incorrect pseudo labels are still prone to accumulate and degrade the performance. To this end, we further propose an advanced self-training framework (namely ST++), that performs selective re-training via prioritizing reliable unlabeled images based on holistic prediction-level stability. Concretely, several model checkpoints are saved in the first stage supervised training, and the discrepancy of their predictions on the unlabeled image serves as a measurement for reliability. Our image-level selection offers holistic contextual information for learning. We demonstrate that it is more suitable for segmentation than common pixel-wise selection. As a result, ST++ further boosts the performance of our ST. Code is available at https://github.com/LiheYoung/ST-PlusPlus.

For agents

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

get_harvested_code_for_paper("2106.05095")
get_code_for_paper("2106.05095")
have("2106.05095")

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

curl https://syntology.ai/api/ran/2106.05095.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.05095.svg)](https://syntology.ai/paper/2106.05095)

Connect an agent — have() is free.