SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.08487 · 2024

Beyond LLaVA-HD: Diving into High-Resolution Large Multimodal Models

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 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, 2 violated it and 0 have no contract verdict; 7 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2406.08487.json · badge: /api/ran/2406.08487.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
yfzhang114/slime canonical (classifier label)
classifier basis: census
9 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
collate_fn Ran
ran; our contract draft was wrong, not the code
yfzhang114/slime/llava/eval/model_vqa_amber.py
code served (permissive licence) · get_code("20e4f665698a3d18")
How it ran

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

import torch

torch.manual_seed(0)

# The collate_fn expects a batch which is a list/iterable of tuples
# Each tuple contains (input_ids, image_tensor, image_size)
# The function unpacks these, stacks input_ids and image_tensors, and returns image_sizes as-is

# Create sample data items
item1 = (
    torch.randint(0, 100, (4,)),  # input_ids: small tensor
    torch.randn(3, 8, 8),          # image_tensor: 3-channel image, 8x8
    (8, 8)                         # image_size: tuple
)

item2 = (
    torch.randint(0, 100, (4,)),  # input_ids: same shape
    torch.randn(3, 8, 8),          # image_tensor: same shape
    (8, 8)                         # image_size: tuple
)

# Create batch as a list of tuples
batch = [item1, item2]

ARGS = {
    "batch": batch
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
evaluate_relaxed_accuracy Ran
ran, honoured the contract we drafted
yfzhang114/slime/llava/eval/model_vqa_chartqa.py
code served (permissive licence) · get_code("6b695143dc1b2b67")
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 numpy as np

# Build entries for evaluate_relaxed_accuracy
entries = [
    {
        'answer': '42.5',
        'annotation': '42.0'  # Will be converted to list
    },
    {
        'answer': '50%',
        'annotation': ['50%', '49.5%']  # Already a list
    },
    {
        'answer': 'hello',
        'annotation': ['HELLO']  # Exact match (case-insensitive)
    },
    {
        'answer': '100.0',
        'annotation': '105.0'  # Within 5% tolerance
    },
    {
        'answer': 'incorrect',
        'annotation': 'different'  # Will not match
    }
]

ARGS = {
    'entries': entries
}
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 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
get_chunk Ran
ran; our contract draft was wrong, not the code
yfzhang114/slime/llava/eval/model_vqa.py
code served (permissive licence) · get_code("42a46570620cd9fa")
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 list, ndim 1, dtype family int; 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 math

def split_list(lst, n):
    """Split a list into n (roughly) equal-sized chunks"""
    chunk_size = math.ceil(len(lst) / n)
    return [lst[i:i + chunk_size] for i in range(0, len(lst), chunk_size)]

def get_chunk(lst, n, k):
    chunks = split_list(lst, n)
    return chunks[k]

# Setup code
lst = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
n = 3
k = 1

ARGS = {
    "lst": lst,
    "n": n,
    "k": k
}
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 list, ndim 1, dtype family int — 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
is_none Ran
ran, violated the contract we drafted
yfzhang114/slime/llava/eval/model_vqa_mmbench.py
pointer only (licence: Apache-2.0); 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: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from ASTRAL-Group/ASTRA, 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.

relaxed_correctness Ran
ran, violated the contract we drafted
yfzhang114/slime/llava/eval/model_vqa_chartqa.py
code served (permissive licence) · get_code("3dd635048bda9fb0")
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 numpy as np

# Set seed for reproducibility
torch.manual_seed(0)

# Define arguments for relaxed_correctness function
ARGS = {
    "target": "100.5",
    "prediction": "102.0",
    "max_relative_change": 0.05
}
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 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
resize_and_pad_image Ran
ran; our contract draft was wrong, not the code
yfzhang114/slime/llava/mm_utils.py
code served (permissive licence) · get_code("468eedeba67f1b00")
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 Image. 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 numpy as np

# Create a sample image for testing
# Create a simple RGB image with distinct dimensions (100x75)
img_array = np.random.randint(0, 256, (75, 100, 3), dtype=np.uint8)
sample_image = Image.fromarray(img_array, mode='RGB')

# Define target resolution (width, height)
target_resolution = (200, 150)

ARGS = {
    "image": sample_image,
    "target_resolution": target_resolution,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
select_best_resolution Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yfzhang114/slime/llava/mm_utils.py
code served (permissive licence) · get_code("3999ff487573f32c")
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. 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

torch.manual_seed(0)

# Construct valid arguments for select_best_resolution
# The function takes:
# 1. original_size: tuple of (width, height)
# 2. possible_resolutions: list of tuples [(width1, height1), (width2, height2), ...]

original_size = (640, 480)  # width=640, height=480

possible_resolutions = [
    (256, 256),
    (512, 512),
    (1024, 768),
    (768, 1024),
    (1280, 960),
]

ARGS = {
    "original_size": original_size,
    "possible_resolutions": possible_resolutions,
}
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 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
select_best_resolution_uhd Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yfzhang114/slime/llava/mm_utils.py
code served (permissive licence) · get_code("b3954d72b4333be4")
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. 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 math

def select_best_resolution_uhd(original_size, processor_size):
    """
    Selects the best resolution from a list of possible resolutions based on the original size.

    Args:
        original_size (tuple): The original size of the image in the format (width, height).
        possible_resolutions (list): A list of possible resolutions in the format [(width1, height1), (width2, height2), ...].

    Returns:
        tuple: The best fit resolution in the format (width, height).
    """
    IMAGE_WIDTH, IMAGE_HEIGHT = processor_size
    origin_image_width, origin_image_height = original_size
    scale = origin_image_width * origin_image_height / (IMAGE_WIDTH * IMAGE_HEIGHT)
    scale = math.ceil(scale)
    if scale > 6:
        scale = 6
    elif scale == 1:
        scale = 2

    def factorize(n):
        factors = []
        for i in range(1, n + 1):
            if n % i == 0:
                factors.append((i / (n / i), i, n // i))
        return factors
    numbers = [1, 2, 3, 4, 5, 6, 7]
    factor_dict = {}
    for num in numbers:
        factor_dict[num] = factorize(num)
    available_ratios = []
    if scale <= 2:
        available_ratios = factor_dict[scale] + factor_dict[scale + 1]
    else:
        available_ratios = factor_dict[scale - 1] + factor_dict[scale] + factor_dict[scale + 1]
    best_fit = None
    max_effective_resolution = 0
    min_wasted_resolution = float('inf')
    for r, w_slice, h_slice in available_ratios:
        width, height = (w_slice * IMAGE_WIDTH, h_slice * IMAGE_HEIGHT)
        scale = min(width / origin_image_width, height / origin_image_height)
        downscaled_width, downscaled_height = (int(origin_image_width * scale), int(origin_image_height * scale))
        effective_resolution = min(downscaled_width * downscaled_height, origin_image_width * origin_image_height)
        wasted_resolution = width * height - effective_resolution
        if effective_resolution > max_effective_resolution or (effective_resolution == max_effective_resolution and wasted_resolution < min_wasted_resolution):
            max_effective_resolution = effective_resolution
            min_wasted_resolution = wasted_resolution
            best_fit = (width, height)
    return best_fit


ARGS = {
    "original_size": (1024, 768),
    "processor_size": (512, 512)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family 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
yfzhang114/slime/llava/eval/model_vqa.py
code served (permissive licence) · get_code("076c252c52cbb161")
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 list. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import math

def split_list(lst, n):
    """Split a list into n (roughly) equal-sized chunks"""
    chunk_size = math.ceil(len(lst) / n)
    return [lst[i:i + chunk_size] for i in range(0, len(lst), chunk_size)]

# Setup code
lst = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
n = 3

ARGS = {
    "lst": lst,
    "n": n
}
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 list — a list output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
unpad_image Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from LLaVA-VL/LLaVA-NeXT
pointer only · get_code("7606525af238fb64")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 9 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 9 the same month, 9 later) and 1 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
select_best_resolution 73 other papers
15 dated earlier, 8 the same month, 50 later · 7 only as an assembly brick
2106.06909 (2021-06)
GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audio
2304.08485 (2023-04)
Visual Instruction Tuning
2308.13566 (2023-08)
MLLM-DataEngine: An Iterative Refinement Approach for MLLM
and 70 more not named here
unpad_image 69 other papers
9 dated earlier, 7 the same month, 53 later · 5 only as an assembly brick
2304.08485 (2023-04)
Visual Instruction Tuning
2309.09958 (2023-09)
An Empirical Study of Scaling Instruct-Tuned Large Multimodal Models
2310.03744 (2023-10)
Improved Baselines with Visual Instruction Tuning
and 66 more not named here
resize_and_pad_image 65 other papers
13 dated earlier, 5 the same month, 47 later
2106.06909 (2021-06)
GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audio
2308.13566 (2023-08)
MLLM-DataEngine: An Iterative Refinement Approach for MLLM
2310.08588 (2023-10)
Octopus: Embodied Vision-Language Programmer from Environmental Feedback
and 62 more not named here
collate_fn 54 other papers
11 dated earlier, 4 the same month, 39 later
2308.13566 (2023-08)
MLLM-DataEngine: An Iterative Refinement Approach for MLLM
2310.08588 (2023-10)
Octopus: Embodied Vision-Language Programmer from Environmental Feedback
2311.07362 (2023-11)
Volcano: Mitigating Multimodal Hallucination through Self-Feedback Guided Revision
and 51 more not named here
evaluate_relaxed_accuracy 6 other papers
1 the same month, 5 later
2406.15334 (2024-06)
Multimodal Task Vectors Enable Many-Shot Multimodal In-Context Learning
2407.01976 (2024-07)
A Bounding Box is Worth One Token: Interleaving Layout and Text in a Large Language Model for Document Understanding
2410.16261 (2024-10)
Mini-InternVL: A Flexible-Transfer Pocket Multimodal Model with 5% Parameters and 90% Performance
and 3 more not named here
relaxed_correctness 6 other papers
1 the same month, 5 later
2406.15334 (2024-06)
Multimodal Task Vectors Enable Many-Shot Multimodal In-Context Learning
2407.01976 (2024-07)
A Bounding Box is Worth One Token: Interleaving Layout and Text in a Large Language Model for Document Understanding
2410.16261 (2024-10)
Mini-InternVL: A Flexible-Transfer Pocket Multimodal Model with 5% Parameters and 90% Performance
and 3 more not named 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

Seeing clearly with high resolution is a foundation of Large Multimodal Models (LMMs), which has been proven to be vital for visual perception and reasoning. Existing works usually employ a straightforward resolution upscaling method, where the image consists of global and local branches, with the latter being the sliced image patches but resized to the same resolution as the former. This means that higher resolution requires more local patches, resulting in exorbitant computational expenses, and meanwhile, the dominance of local image tokens may diminish the global context. In this paper, we dive into the problems and propose a new framework as well as an elaborate optimization strategy. Specifically, we extract contextual information from the global view using a mixture of adapters, based on the observation that different adapters excel at different tasks. With regard to local patches, learnable query embeddings are introduced to reduce image tokens, the most important tokens accounting for the user question will be further selected by a similarity-based selector. Our empirical results demonstrate a `less is more' pattern, where \textit{utilizing fewer but more informative local image tokens leads to improved performance}. Besides, a significant challenge lies in the training strategy, as simultaneous end-to-end training of the global mining block and local compression block does not yield optimal results. We thus advocate for an alternating training way, ensuring balanced learning between global and local aspects. Finally, we also introduce a challenging dataset with high requirements for image detail, enhancing the training of the local compression layer. The proposed method, termed LMM with Sophisticated Tasks, Local image compression, and Mixture of global Experts (SliME), achieves leading performance across various benchmarks with only 2 million training data.

For agents

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

get_harvested_code_for_paper("2406.08487")
get_code_for_paper("2406.08487")
have("2406.08487")

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

curl https://syntology.ai/api/ran/2406.08487.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.08487.svg)](https://syntology.ai/paper/2406.08487)

Connect an agent — have() is free.