SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2211.09233 · 2022

Prompt Tuning for Parameter-efficient Medical Image Segmentation

arXiv · PDF · Open in the Atlas

Code that ran

12 of 16 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 12 have no contract verdict; 0 failures of our instrument; 4 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/2211.09233.json · badge: /api/ran/2211.09233.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
marcdcfischer/punet canonical (classifier label)
classifier basis not recorded on the link
4 ran · 2 no recorded run
marcdcfischer/punetr community (classifier label)
classifier basis not recorded on the link
8 ran · 2 no recorded run
FunctionStatusWhere it lives
add_callback_specific_args Ran
ran on a synthesized input
marcdcfischer/punet/src/utils/callbacks.py
code served (permissive licence) · get_code("55f4926388207295")
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 ArgumentParser. 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 argparse import ArgumentParser

# Create a minimal parent parser
parent_parser = ArgumentParser(add_help=False)
parent_parser.add_argument('--some_arg', type=str, default='default')

ARGS = {
    'parent_parser': parent_parser
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
from_argparse_args Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/old_lightning_argparse.py
code served (permissive licence) · get_code("fae5b0a45b8473ed")
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 MockTrainer. 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
from argparse import ArgumentParser, Namespace
from unittest.mock import MagicMock

torch.manual_seed(0)

# Create a minimal mock LightningDataModule class
class MockLightningDataModule:
    def __init__(self, batch_size: int = 32, num_workers: int = 0, **kwargs):
        self.batch_size = batch_size
        self.num_workers = num_workers
        for key, value in kwargs.items():
            setattr(self, key, value)
    
    @classmethod
    def add_argparse_args(cls, parser):
        parser.add_argument('--batch_size', type=int, default=32)
        parser.add_argument('--num_workers', type=int, default=0)
        return parser
    
    @classmethod
    def parse_argparser(cls, args):
        return args

# Create a minimal mock Trainer class
class MockTrainer:
    def __init__(self, max_epochs: int = 10, learning_rate: float = 0.001, **kwargs):
        self.max_epochs = max_epochs
        self.learning_rate = learning_rate
        for key, value in kwargs.items():
            setattr(self, key, value)
    
    @classmethod
    def add_argparse_args(cls, parser):
        parser.add_argument('--max_epochs', type=int, default=10)
        parser.add_argument('--learning_rate', type=float, default=0.001)
        return parser
    
    @classmethod
    def parse_argparser(cls, args):
        return args

# Test with ArgumentParser
parser = ArgumentParser(add_help=False)
parser.add_argument('--max_epochs', type=int, default=5)
parser.add_argument('--learning_rate', type=float, default=0.01)
parser.add_argument('--logger', type=bool, default=False)

parsed_args = parser.parse_args([])
namespace_args = Namespace(max_epochs=5, learning_rate=0.01, logger=False)

ARGS = {
    'cls': MockTrainer,
    'args': namespace_args,
    'kwargs': {'logger': 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 MockTrainer — a MockTrainer output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockTrainer 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
list_objects Ran
ran on a synthesized input
marcdcfischer/punet/src/utils/aws.py
code served (permissive licence) · get_code("3162d297bfe0a0b3")
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, 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 logging
from unittest.mock import Mock, MagicMock
from botocore.exceptions import ClientError

# Create a mock S3 bucket object with the required attributes and methods
mock_bucket = Mock()
mock_bucket.name = "test-bucket"

# Create mock S3 objects with key attributes
mock_object_1 = Mock()
mock_object_1.key = "file1.txt"

mock_object_2 = Mock()
mock_object_2.key = "prefix/file2.txt"

mock_object_3 = Mock()
mock_object_3.key = "prefix/file3.txt"

# Setup the bucket.objects.all() method to return all objects
mock_bucket.objects.all.return_value = [mock_object_1, mock_object_2, mock_object_3]

# Setup the bucket.objects.filter() method to return filtered objects
mock_bucket.objects.filter.return_value = [mock_object_2, mock_object_3]

# Define arguments for the function
ARGS = {
    "bucket": mock_bucket,
    "prefix": None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
botocore — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list 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
parse_argparser Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/old_lightning_argparse.py
code served (permissive licence) · get_code("78ad1521c16f7df8")
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 Namespace. 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 inspect
from argparse import ArgumentParser, Namespace
import torch
import torch.nn as nn

torch.manual_seed(0)

# Create mock Lightning classes for testing
class MockLightningDataModule:
    def __init__(
        self,
        batch_size: int = 32,
        num_workers: int = 0,
        shuffle: bool = True,
        pin_memory: bool = False,
        timeout: float = 0.0,
    ):
        self.batch_size = batch_size
        self.num_workers = num_workers
        self.shuffle = shuffle
        self.pin_memory = pin_memory
        self.timeout = timeout

class MockTrainer:
    def __init__(
        self,
        max_epochs: int = 10,
        accelerator: str = "cpu",
        devices: int = 1,
        enable_checkpointing: bool = True,
        logger: bool = True,
        log_every_n_steps: int = 50,
    ):
        self.max_epochs = max_epochs
        self.accelerator = accelerator
        self.devices = devices
        self.enable_checkpointing = enable_checkpointing
        self.logger = logger
        self.log_every_n_steps = log_every_n_steps

# Create test argument parser
arg_parser = ArgumentParser()
arg_parser.add_argument('--batch_size', type=int, default=32)
arg_parser.add_argument('--num_workers', type=int, default=0)
arg_parser.add_argument('--shuffle', type=bool, default=True)
arg_parser.add_argument('--pin_memory', type=bool, default=False)
arg_parser.add_argument('--timeout', type=float, default=0.0)

ARGS = {
    'cls': MockLightningDataModule,
    'arg_parser': arg_parser
}
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 Namespace — a Namespace output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Namespace 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
parse_env_variables Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/old_lightning_argparse.py
code served (permissive licence) · get_code("5b13add0dd978d7d")
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 Namespace. 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
import inspect
from argparse import Namespace
from typing import Type, Union

# Mock Lightning classes for testing
class MockLightningDataModule:
    def __init__(self, batch_size: int = 32, num_workers: int = 0):
        self.batch_size = batch_size
        self.num_workers = num_workers

class MockTrainer:
    def __init__(
        self,
        max_epochs: int = 10,
        devices: Union[int, str] = 1,
        accelerator: str = 'cpu',
        precision: Union[int, str] = 32,
        enable_checkpointing: bool = True,
        logger: bool = True,
        log_every_n_steps: int = 50
    ):
        self.max_epochs = max_epochs
        self.devices = devices
        self.accelerator = accelerator
        self.precision = precision
        self.enable_checkpointing = enable_checkpointing
        self.logger = logger
        self.log_every_n_steps = log_every_n_steps

# Create test class for demonstration
test_cls = MockTrainer

ARGS = {
    'cls': test_cls,
    'template': 'PL_%(cls_name)s_%(cls_argument)s'
}
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 Namespace — a Namespace output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Namespace 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
setup_checkpointing Ran
ran on a synthesized input
marcdcfischer/punet/src/utils/callbacks.py
code served (permissive licence) · get_code("b8bf5545c88f7ce3")
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 ModelCheckpoint. 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 pathlib as plb
import pytorch_lightning as pl
import os
import tempfile

class MockHParams:
    def __init__(self):
        self.s3_bucket = ''
        self.online_on = False
        self.run_name = 'test_run'
        self.default_root_dir = tempfile.gettempdir()
        self.ckpt_save_last = True
        self.loss_weight_segmentation = 1.0
        self.__dict__ = self.__dict__

hparams = MockHParams()

ARGS = {
    'hparams': hparams,
    'save_top_k': 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pytorch_lightning — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ModelCheckpoint — a ModelCheckpoint output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ModelCheckpoint 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
setup_checkpointing Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/callbacks.py
code served (permissive licence) · get_code("a442905baf491e84")
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 ModelCheckpoint. 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 pathlib as plb
import pytorch_lightning as pl

class HParams:
    def __init__(self):
        self.s3_bucket = ''
        self.online_off = False
        self.run_name = 'test_run'
        self.default_root_dir = '/tmp/lightning_logs'
        self.ckpt_save_last = True
        self.loss_weight_segmentation = 1.0
        self.__dict__ = self.__dict__

hparams = HParams()

ARGS = {
    'hparams': hparams,
    'save_top_k': 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pytorch_lightning — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ModelCheckpoint — a ModelCheckpoint output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ModelCheckpoint 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
setup_loggers Ran
ran on a synthesized input
marcdcfischer/punet/src/utils/logging_custom.py
code served (permissive licence) · get_code("ebea7d126a7d1a2f")
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, 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 tempfile
import pathlib as plb

class HParams:
    def __init__(self):
        self.online_logger = 'wandb'
        self.online_project = 'test_project'
        self.tags = ['test', 'tag']
        self.online_entity = 'test_entity'
        self.online_on = False
        self.default_root_dir = None

hparams = HParams()

with tempfile.TemporaryDirectory() as tmpdir:
    path_root = tmpdir

ARGS = {
    'hparams': hparams,
    'path_root': path_root
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pytorch_lightning — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list 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
stable_softmax Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/stable_softmax.py
code served (permissive licence) · get_code("e38ff7174bf4d914")
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 2, dtype family float, shape (2, 4) 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

torch.manual_seed(0)

# Create a small logits tensor with distinct dimensions for probing
# Shape: (batch=2, features=4)
logits = torch.randn(2, 4)

# Temperature parameter for scaling
temperature = 1.0

# Dimension to apply softmax over (can be int or Sequence of ints)
dim = 1

ARGS = {
    "logits": logits,
    "temperature": temperature,
    "dim": dim,
}
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 2, dtype family float, shape (2, 4) — shape (2, 4) 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
str_to_bool Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/old_lightning_parsing.py
code served (permissive licence) · get_code("e5cfe85a69313b6f")
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 bool, 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

# The function str_to_bool takes a single parameter 'val' which is a string
# It converts string representations of truth values to boolean
# Looking at the function body, it:
# 1. Calls .lower() on the input string
# 2. Checks if the lowercased string is in specific tuples
# 3. Returns True, False, or the original string value

# Test with a valid true value
ARGS = {
    "val": "true"
}
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 bool, 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
str_to_bool_or_int Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/old_lightning_parsing.py
code served (permissive licence) · get_code("5c7789af09eb566b")
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 bool, 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.

# Analysis: str_to_bool_or_int takes a single parameter `val` of type str
# The function converts string representations to bool, int, or returns the string

ARGS = {
    "val": "true"
}
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 bool, 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
str_to_bool_or_str Ran
ran on a synthesized input
marcdcfischer/punetr/src/utils/old_lightning_parsing.py
code served (permissive licence) · get_code("672c119bf3562487")
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 bool, 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

def str_to_bool_or_str(val: str) -> bool | str:
    """Possibly convert a string representation of truth to bool. Returns the input otherwise. Based on the python
    implementation distutils.utils.strtobool.
    True values are 'y', 'yes', 't', 'true', 'on', and '1'; false values are 'n', 'no', 'f', 'false', 'off', and '0'.
    """
    lower = val.lower()
    if lower in ('y', 'yes', 't', 'true', 'on', '1'):
        return True
    if lower in ('n', 'no', 'f', 'false', 'off', '0'):
        return False
    return val

ARGS = {
    "val": "true"
}
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 bool, 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
download_with_default_configuration Not yet run
unverified (harvested, no recorded run)
marcdcfischer/punet/src/utils/aws.py
code served (permissive licence) · get_code("11e5b48f56f056e5")
How it ran

no invocation recorded.

fetch_ckpt Not yet run
unverified (harvested, no recorded run)
marcdcfischer/punet/src/utils/aws.py
code served (permissive licence) · get_code("d2e26bc2ed5c46ec")
How it ran

no invocation recorded.

fetch_ckpt Not yet run
unverified (harvested, no recorded run)
marcdcfischer/punetr/src/utils/aws.py
code served (permissive licence) · get_code("60d28b11dc70aefc")
How it ran

no invocation recorded.

setup_loggers Not yet run
unverified (harvested, no recorded run)
marcdcfischer/punetr/src/utils/logging_custom.py
code served (permissive licence) · get_code("d5b89e699579032c")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 16 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.

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

Neural networks pre-trained on a self-supervision scheme have become the standard when operating in data rich environments with scarce annotations. As such, fine-tuning a model to a downstream task in a parameter-efficient but effective way, e.g. for a new set of classes in the case of semantic segmentation, is of increasing importance. In this work, we propose and investigate several contributions to achieve a parameter-efficient but effective adaptation for semantic segmentation on two medical imaging datasets. Relying on the recently popularized prompt tuning approach, we provide a prompt-able UNet (PUNet) architecture, that is frozen after pre-training, but adaptable throughout the network by class-dependent learnable prompt tokens. We pre-train this architecture with a dedicated dense self-supervision scheme based on assignments to online generated prototypes (contrastive prototype assignment, CPA) of a student teacher combination alongside a concurrent segmentation loss on a subset of classes. We demonstrate that the resulting neural network model is able to attenuate the gap between fully fine-tuned and parameter-efficiently adapted models on CT imaging datasets. As such, the difference between fully fine-tuned and prompt-tuned variants amounts to only 3.83 pp for the TCIA/BTCV dataset and 2.67 pp for the CT-ORG dataset in the mean Dice Similarity Coefficient (DSC, in %) while only prompt tokens, corresponding to 0.85% of the pre-trained backbone model with 6.8M frozen parameters, are adjusted. The code for this work is available on https://github.com/marcdcfischer/PUNet .

For agents

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

get_harvested_code_for_paper("2211.09233")
get_code_for_paper("2211.09233")
have("2211.09233")

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

curl https://syntology.ai/api/ran/2211.09233.json

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

[![Syntology run record](https://syntology.ai/api/ran/2211.09233.svg)](https://syntology.ai/paper/2211.09233)

Connect an agent — have() is free.