SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.11280 · 2024

ISR-DPO: Aligning Large Multimodal Models for Videos by Iterative Self-Retrospective DPO

arXiv · PDF · Open in the Atlas

Code that ran

26 of 30 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; 5 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 19 failures of our instrument; 4 no recorded run); 3 of them were also run on the host, not in the jail, by the peer-agreement lane that set their level, 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/2406.11280.json · badge: /api/ran/2406.11280.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
yonseivnl/vlm-rlaif canonical (classifier label)
classifier basis: census
12 ran · 3 no recorded run
snumprlab/SRT canonical (classifier label)
classifier basis: census
12 ran · 1 no recorded run
snumprlab/isr-dpo canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
FunctionStatusWhere it lives
batch_select Ran
ran; our contract draft was wrong, not the code
yonseivnl/vlm-rlaif/RLAIF/models/reward_model.py
code served (permissive licence) · get_code("2154e44a6b59e432")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 2, dtype family int, shape (2, 3) as a later run recorded it. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T22:51:12Z). 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 input tensor: batch_size=2, num_features=4
input = torch.tensor([
    [0, 1, 2, 5],
    [3, 0, 9, 4],
])

# Create index tensor: batch_size=2, num_selections=3
# Each row contains indices to select from the corresponding input row
index = torch.tensor([
    [0, 1, 3],
    [1, 0, 2],
])

ARGS = {
    "input": input,
    "index": index,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 int, shape (2, 3) — shape (2, 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
get_id_from_path Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/inference/inference.py
pointer only (licence: NONE); the stored copy get_code serves, from RifleZhang/LLaVA-Hound-DPO, is pointer only (licence: NONE) · get_code("f1a2ea781d07a5ee")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

unpack_dict Ran
ran; our contract draft was wrong, not the code
yonseivnl/vlm-rlaif/RLAIF/models/reward_model.py
code served (permissive licence) · get_code("a6989412aca721c1")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned tuple. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-05T01:20:33Z). 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)

# Create a sample dictionary with various keys and values
d = {
    'key1': torch.randn(2, 4),
    'key2': torch.randn(3, 5),
    'key3': 'value3',
    'key4': 42,
    'key5': [1, 2, 3],
}

# Define keys to extract
keys = ['key1', 'key3', 'key4']

# Define return type
return_type = tuple

ARGS = {
    'd': d,
    'keys': keys,
    'return_type': return_type,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
all_gather_and_cat Ran
ran, honoured the contract we drafted
yonseivnl/vlm-rlaif/RLAIF/models/distributed_utils.py
code served (permissive licence) · get_code("5aa59c85970b28a7")
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 os
import torch
import torch.distributed as dist

torch.manual_seed(0)

# Set up a non-distributed environment (WORLD_SIZE = 1)
# This allows the function to run without actual distributed setup
os.environ['WORLD_SIZE'] = '1'

# Create a small tensor for testing
# Shape: (2, 4) - batch size 2, feature dimension 4
tensor = torch.randn(2, 4)

ARGS = {
    'tensor': tensor,
    'dim': 0
}
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
alleq Ran
ran; our contract draft was wrong, not the code
yonseivnl/vlm-rlaif/RLAIF/data_utils/common_utils.py
code served (permissive licence) · get_code("d991ae45245789f9")
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
import torch.nn as nn
from typing import Callable, Optional, Sequence

torch.manual_seed(0)

# Test case 1: List of equal integers
l1 = [5, 5, 5, 5]
f1 = lambda x, y: x == y

# Test case 2: List of equal tensors
l2 = [torch.tensor([1, 2, 3]), torch.tensor([1, 2, 3]), torch.tensor([1, 2, 3])]
f2 = lambda x, y: torch.equal(x, y)

# Test case 3: List of equal strings
l3 = ["hello", "hello", "hello"]
f3 = lambda x, y: x == y

# Test case 4: List with single element (edge case)
l4 = [42]
f4 = lambda x, y: x == y

# Test case 5: Default comparison function with equal elements
l5 = [1, 1, 1]

ARGS = {
    "l": l1,
    "f": f1,
}
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
cast_with_native_amp Ran
ran; our contract draft was wrong, not the code
yonseivnl/vlm-rlaif/RLAIF/models/rl_trainer.py
code served (permissive licence) · get_code("af2735cffc6ee536")
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 function. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 typing import Callable, Optional

torch.manual_seed(0)

# Create a simple function that will be wrapped
def simple_forward(x: torch.Tensor) -> torch.Tensor:
    """A simple function that takes a tensor and returns a tensor."""
    return x * 2

# Test with different mixed precision modes
ARGS = {
    "func": simple_forward,
    "mixed_precision": "fp16"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
accelerate, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
convert_to_scalar Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/trl/core.py
pointer only (licence: NONE) · get_code("c32b636f13ed9ee5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

create_scheduler Ran
ran; our contract draft was wrong, not the code
yonseivnl/vlm-rlaif/RLAIF/models/trainer_utils.py
code served (permissive licence) · get_code("42596ac916e883a5")
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 LambdaLR. 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 torch.optim import Adam

torch.manual_seed(0)

# Create a minimal optimizer
model = nn.Linear(4, 3)
optimizer = Adam(model.parameters(), lr=1e-4)

# Create a minimal args object with required attributes and methods
class Args:
    def __init__(self):
        self.lr_scheduler_type = "linear"
    
    def get_warmup_steps(self, num_training_steps):
        return int(0.1 * num_training_steps)

args = Args()

# Set up the remaining parameters
num_training_steps = 1000
lr_scheduler = None

ARGS = {
    "args": args,
    "optimizer": optimizer,
    "lr_scheduler": lr_scheduler,
    "num_training_steps": num_training_steps,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type LambdaLR — a LambdaLR output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a LambdaLR 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
flatten_dict Ran
ran; our contract draft was wrong, not the code
yonseivnl/vlm-rlaif/RLAIF/data_utils/common_utils.py
code served (permissive licence) · get_code("f5ff77db7757bc92")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Construct a nested dictionary with various types of values
nested_dict = {
    'model': {
        'layer1': {
            'weights': torch.randn(2, 4),
            'bias': torch.randn(4)
        },
        'layer2': {
            'weights': torch.randn(3, 4)
        }
    },
    'config': {
        'learning_rate': 0.001,
        'batch_size': 32
    },
    'data': 42
}

# Define a postprocess function that processes values
def my_postprocess(v):
    if isinstance(v, torch.Tensor):
        return v.sum().item()
    return v

ARGS = {
    'nested': nested_dict,
    'sep': '.',
    'postprocess_fn': my_postprocess
}
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 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
get_chunk Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/llava/eval/model_vqa.py
pointer only (licence: NONE) · get_code("42a46570620cd9fa")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

get_id_from_frame_path Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/data_processing/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from riflezhang/llava-hound-dpo, is pointer only (licence: NONE) · get_code("f345a47ea5dbb5bc")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

get_image Ran
ran, honoured the contract we drafted
snumprlab/SRT/data_processing/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from riflezhang/llava-hound-dpo, is pointer only (licence: NONE) · get_code("1f804baed33a101d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

get_last_checkpoint Ran
ran; our contract draft was wrong, not the code
yonseivnl/vlm-rlaif/RLAIF/lora_utils.py
code served (permissive licence) · get_code("89ce33341af66037")
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) 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 os
import tempfile
import shutil

# Create a temporary directory structure for testing
temp_dir = tempfile.mkdtemp()

# Create a checkpoint directory with some checkpoint subdirectories
checkpoint_dir = os.path.join(temp_dir, 'checkpoints')
os.makedirs(checkpoint_dir, exist_ok=True)

# Create some checkpoint directories
os.makedirs(os.path.join(checkpoint_dir, 'checkpoint-100'), exist_ok=True)
os.makedirs(os.path.join(checkpoint_dir, 'checkpoint-200'), exist_ok=True)
os.makedirs(os.path.join(checkpoint_dir, 'checkpoint-50'), exist_ok=True)

ARGS = {
    'checkpoint_dir': checkpoint_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
get_score Ran
ran, honoured the contract we drafted
snumprlab/SRT/inference/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from riflezhang/llava-hound-dpo, is pointer only (licence: NONE) · get_code("5321a2aa0d7681cb")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

is_none Ran
ran, violated the contract we drafted
snumprlab/SRT/llava/eval/model_vqa_mmbench.py
pointer only (licence: NONE); the stored copy get_code serves, from ASTRAL-Group/ASTRA, is pointer only (licence: NONE) · get_code("bae18947b56f2be1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

load_frames Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/data_processing/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from riflezhang/llava-hound-dpo, is pointer only (licence: NONE) · get_code("6fac133e5d5b321b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

make_data Ran
ran; our contract draft was wrong, not the code
snumprlab/isr-dpo/Evaluation/eval/eval_qa.py
pointer only (licence: NONE) · get_code("63c0b0c7b99fe5b6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/isr-dpo 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.

maybe_truncate Ran
ran; our contract draft was wrong, not the code
snumprlab/isr-dpo/Evaluation/eval/eval_qa.py
pointer only (licence: NONE) · get_code("0f72ca36a1580010")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/isr-dpo 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.

openai_generate Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/inference/chatgpt_inference.py
pointer only (licence: NONE) · get_code("ddd3b7e8f7827aba")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

remove_pad_and_left_pad Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yonseivnl/vlm-rlaif/RLAIF/models/ppo_trainer.py
code served (permissive licence) · get_code("606e2b3506d6ed7d")
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 int, 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 batch of completions with some padding tokens
# Shape: (batch_size=2, sequence_length=4)
completions = torch.tensor([
    [1, 2, 0, 0],      # First completion with padding
    [3, 4, 5, 0]       # Second completion with padding
], dtype=torch.long)

pad_token_id = 0

ARGS = {
    "completions": completions,
    "pad_token_id": pad_token_id
}
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 int, 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
remove_special_tokens Ran
ran, violated the contract we drafted
snumprlab/SRT/inference/inference.py
pointer only (licence: NONE); the stored copy get_code serves, from RifleZhang/LLaVA-Hound-DPO, is pointer only (licence: NONE) · get_code("4354061aab23f5b9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

setup Ran
ran, honoured the contract we drafted
yonseivnl/vlm-rlaif/RLAIF/models/distributed_utils.py
code served (permissive licence) · get_code("4098e9c2afde4a5d")
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 int; no output shape is stated. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 os
import sys
import torch
import torch.distributed as dist

torch.manual_seed(0)

# The setup function takes optional rank and world_size parameters
# If None, it tries to read from environment variables RANK and WORLD_SIZE
# For testing, we'll provide explicit values to avoid needing env vars

ARGS = {
    'rank': 0,
    'world_size': 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 int — 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
split_list Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/llava/eval/model_vqa.py
pointer only (licence: NONE) · get_code("076c252c52cbb161")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

stack_dicts Ran
ran; our contract draft was wrong, not the code
snumprlab/SRT/trl/core.py
pointer only (licence: NONE) · get_code("085e70e0bd6e492d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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.

unpack_dict Ran
ran, honoured the contract we drafted
yonseivnl/vlm-rlaif/RLAIF/data_utils/common_utils.py
code served (permissive licence) · get_code("e5295408ef50f6e8")
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 2, 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 numpy as np
from typing import Dict, Sequence, Union

torch.manual_seed(0)

# Construct the arguments for unpack_dict
d = {
    'key1': torch.tensor([1.0, 2.0, 3.0, 4.0]),
    'key2': torch.tensor([5.0, 6.0, 7.0, 8.0]),
    'key3': np.array([9, 10, 11, 12]),
    'key4': 'some_value'
}

keys = ['key1', 'key2', 'key3']

return_type = tuple

ARGS = {
    'd': d,
    'keys': keys,
    'return_type': return_type
}
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 tuple, ndim 2, 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
whiten Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yonseivnl/vlm-rlaif/RLAIF/models/ppo_trainer.py
code served (permissive licence) · get_code("9a0b92f09a2d6c76")
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 (8, 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 tensor for testing
# The function expects values to be a tensor
# For 'full_batch' and 'per_gpu' modes, it needs at least 8 elements in batch dimension
values = torch.randn(8, 4)

ARGS = {
    'values': values,
    'shift_mean': True,
    'epsilon': 1e-08,
    'async_stats': 'none'
}
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 (8, 4) — shape (8, 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
flatten_dict Not yet run
unverified (harvested, no recorded run)
snumprlab/SRT/trl/core.py
pointer only (licence: NONE) · get_code("a51ae45a1f6c0b5b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, snumprlab/SRT 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. No invocation recorded.

get_peft_state_maybe_zero_3 Not yet run
unverified (harvested, no recorded run)
yonseivnl/vlm-rlaif/RLAIF/finetune_policy_init.py
code served (permissive licence) · get_code("f0eed88e133ab3b9")
How it ran

no invocation recorded.

get_peft_state_non_lora_maybe_zero_3 Not yet run
unverified (harvested, no recorded run)
yonseivnl/vlm-rlaif/RLAIF/finetune_policy_init.py
code served (permissive licence) · get_code("1c53657305b66e9f")
How it ran

no invocation recorded.

maybe_zero_3 Not yet run
unverified (harvested, no recorded run)
yonseivnl/vlm-rlaif/RLAIF/finetune_policy_init.py
code served (permissive licence) · get_code("616ffbdc154ed2d8")
How it ran

no invocation recorded.

Also present in other papers' code

Of 30 samples lifted from this paper's linked repositories, 30 carry a function identity and 0 do not. Of the 29 distinct functions with an identity, 24 are also present in other papers whose code we hold (18 with at least one paper dated earlier, 12 the same month, 18 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
split_list 197 other papers
60 dated earlier, 22 the same month, 108 later, 6 same year, order not determinable, 1 undated
2112.05682 (2021-12)
Self-attention Does Not Need $O(n^2)$ Memory
2306.17107 (2023-06)
LLaVAR: Enhanced Visual Instruction Tuning for Text-Rich Image Understanding
2307.10490 (2023-07)
Abusing Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs
and 194 more not named here
get_chunk 189 other papers
58 dated earlier, 21 the same month, 104 later, 6 same year, order not determinable
2112.05682 (2021-12)
Self-attention Does Not Need $O(n^2)$ Memory
2306.17107 (2023-06)
LLaVAR: Enhanced Visual Instruction Tuning for Text-Rich Image Understanding
2307.10490 (2023-07)
Abusing Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs
and 186 more not named here
is_none 83 other papers
28 dated earlier, 10 the same month, 43 later, 2 same year, order not determinable
2112.05682 (2021-12)
Self-attention Does Not Need $O(n^2)$ Memory
2308.10253 (2023-08)
StableLLaVA: Enhanced Visual Instruction Tuning with Synthesized Image-Dialogue Data
2308.13566 (2023-08)
MLLM-DataEngine: An Iterative Refinement Approach for MLLM
and 80 more not named here
maybe_zero_3 37 other papers
13 dated earlier, 4 the same month, 20 later
2310.01779 (2023-10)
HallE-Control: Controlling Object Hallucination in Large Multimodal Models
2311.16839 (2023-11)
Beyond Hallucinations: Enhancing LVLMs through Hallucination-Aware Direct Preference Optimization
2312.02949 (2023-12)
LLaVA-Grounding: Grounded Visual Chat with Large Multimodal Models
and 34 more not named here
get_peft_state_non_lora_maybe_zero_3 23 other papers
7 dated earlier, 3 the same month, 13 later
2310.01779 (2023-10)
HallE-Control: Controlling Object Hallucination in Large Multimodal Models
2312.02949 (2023-12)
LLaVA-Grounding: Grounded Visual Chat with Large Multimodal Models
2401.02330 (2024-01)
LLaVA-Phi: Efficient Multi-Modal Assistant with Small Language Model
and 20 more not named here
convert_to_scalar 18 other papers
1 dated earlier, 1 the same month, 16 later
2305.18290 (2023-05)
Direct Preference Optimization: Your Language Model is Secretly a Reward Model
2406.19371 (2024-06)
Suri: Multi-constraint Instruction Following for Long-form Text Generation
2410.06101 (2024-10)
Coevolving with the Other You: Fine-Tuning LLM with Sequential Cooperative Multi-Agent Reinforcement Learning
and 15 more not named here
flatten_dict 18 other papers
1 dated earlier, 1 the same month, 16 later
2305.18290 (2023-05)
Direct Preference Optimization: Your Language Model is Secretly a Reward Model
2406.19371 (2024-06)
Suri: Multi-constraint Instruction Following for Long-form Text Generation
2410.06101 (2024-10)
Coevolving with the Other You: Fine-Tuning LLM with Sequential Cooperative Multi-Agent Reinforcement Learning
and 15 more not named here
get_last_checkpoint 5 other papers
3 dated earlier, 2 later
2305.14314 (2023-05)
QLORA: Efficient Finetuning of Quantized LLMs
2310.05910 (2023-10)
SALMON: Self-Alignment with Instructable Reward Models
2405.11165 (2024-05)
Automated Multi-level Preference for MLLMs
and 2 more not named here
create_scheduler 4 other papers
2 dated earlier, 1 the same month, 1 later
2309.09055 (2023-09)
Exploring the impact of low-rank adaptation on the performance, efficiency, and regularization of RLHF
2310.05910 (2023-10)
SALMON: Self-Alignment with Instructable Reward Models
2406.12030 (2024-06)
SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model
and 1 more not named here
flatten_dict 4 other papers
3 dated earlier, 1 later
2309.09055 (2023-09)
Exploring the impact of low-rank adaptation on the performance, efficiency, and regularization of RLHF
2310.11564 (2023-10)
Personalized Soups: Personalized Large Language Model Alignment via Post-hoc Parameter Merging
2402.05133 (2024-02)
Personalized Language Modeling from Personalized Human Feedback
and 1 more not named here
get_id_from_path 4 other papers
1 dated earlier, 3 later
2404.01258 (2024-04)
Direct Preference Optimization of Video Large Multimodal Models from Language Model Reward
2410.16198 (2024-10)
Improve Vision Language Model Chain-of-thought Reasoning
2410.17637 (2024-10)
MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models
and 1 more not named here
remove_special_tokens 4 other papers
1 dated earlier, 3 later
2404.01258 (2024-04)
Direct Preference Optimization of Video Large Multimodal Models from Language Model Reward
2410.16198 (2024-10)
Improve Vision Language Model Chain-of-thought Reasoning
2410.17637 (2024-10)
MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models
and 1 more not named here
all_gather_and_cat 3 other papers
2 dated earlier, 1 the same month
2310.05910 (2023-10)
SALMON: Self-Alignment with Instructable Reward Models
2405.11165 (2024-05)
Automated Multi-level Preference for MLLMs
2406.12030 (2024-06)
SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model
batch_select 3 other papers
2 dated earlier, 1 the same month
2310.05910 (2023-10)
SALMON: Self-Alignment with Instructable Reward Models
2403.09472 (2024-03)
Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision
2406.12030 (2024-06)
SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model
setup 3 other papers
2 dated earlier, 1 the same month
2310.05910 (2023-10)
SALMON: Self-Alignment with Instructable Reward Models
2405.11165 (2024-05)
Automated Multi-level Preference for MLLMs
2406.12030 (2024-06)
SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model
unpack_dict 3 other papers
2 dated earlier, 1 the same month
2310.05910 (2023-10)
SALMON: Self-Alignment with Instructable Reward Models
2403.09472 (2024-03)
Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision
2406.12030 (2024-06)
SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model
alleq 2 other papers
2 later
2409.07431 (2024-09)
Synthetic continued pretraining
2603.22213 (2026-03)
SPA: A Simple but Tough-to-Beat Baseline for Knowledge Injection
stack_dicts 2 other papers
2 later
2410.16198 (2024-10)
Improve Vision Language Model Chain-of-thought Reasoning
2410.17637 (2024-10)
MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models
cast_with_native_amp 1 other paper
1 dated earlier
2310.05910 (2023-10)
SALMON: Self-Alignment with Instructable Reward Models
get_id_from_frame_path 1 other paper
1 later
2410.16198 (2024-10)
Improve Vision Language Model Chain-of-thought Reasoning

Shown: the 20 functions present in the most other papers. 4 more shared functions are not listed here.

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

Iterative self-improvement, a concept extending beyond personal growth, has found powerful applications in machine learning, particularly in transforming weak models into strong ones. While recent advances in natural language processing have shown its efficacy through iterative preference optimization, applying this approach to Video Large Multi-modal Models (VLMMs) remains challenging due to modality misalignment. VLMMs struggle with this misalignment during iterative preference modeling, as the self-judge model often prioritizes linguistic knowledge over visual information. Additionally, iterative preference optimization can lead to visually hallucinated verbose responses due to length bias within the self-rewarding cycle. To address these issues, we propose Iterative Self-Retrospective Direct Preference Optimization (ISR-DPO), a method that uses self-retrospection to enhance preference modeling. This approach enhances the self-judge's focus on informative video regions, resulting in more visually grounded preferences. In extensive empirical evaluations across diverse video question answering benchmarks, the ISR-DPO significantly outperforms the state of the art. We are committed to open-sourcing our code, models, and datasets to encourage further investigation.

For agents

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

get_harvested_code_for_paper("2406.11280")
get_code_for_paper("2406.11280")
have("2406.11280")

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

curl https://syntology.ai/api/ran/2406.11280.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.11280.svg)](https://syntology.ai/paper/2406.11280)

Connect an agent — have() is free.