SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.11403 · ICCV · 2023

eP-ALM: Efficient Perceptual Augmentation of Language Models

Matthieu Cord, Mustafa Shukor, Corentin Dancette

arXiv · PDF · Open in the Atlas

Code that ran

15 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 4 have no contract verdict; 11 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2303.11403.json · badge: /api/ran/2303.11403.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
mshukor/eP-ALM canonical (classifier label)
classifier basis not recorded on the link
15 ran · 1 no recorded run
FunctionStatusWhere it lives
all_gather Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/utils.py
code served (permissive licence) · get_code("09ff1aa1042d3b3f")
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. 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 torch
import torch.distributed as dist
import pickle
import logging

torch.manual_seed(0)

# Set up minimal distributed environment (non-initialized)
# The function handles the case where dist is not initialized
# by returning early with [data]

# Create some picklable test data
test_data = {
    "tensor": torch.randn(2, 4),
    "list": [1, 2, 3],
    "string": "hello world"
}

# Since distributed is not initialized in this sandbox,
# the all_gather function will return [data] immediately
# at the first check: if get_world_size() == 1: return [data]
# So we just need to provide valid picklable data

ARGS = {
    "data": test_data,
    "group": 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 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
basic_clean Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/models/clip.py
pointer only (licence: MIT); the stored copy get_code serves, from ltpwy/MSCI, is pointer only (licence: NONE) · get_code("98f385d847636a3e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ltpwy/MSCI, 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.

checkpoint_filter_fn Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/models/vit.py
pointer only (licence: MIT); the stored copy get_code serves, from JiahuaDong/HFC, is pointer only (licence: NONE) · get_code("1f65552c7dfe6207")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from JiahuaDong/HFC, 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.

compute_acc Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mshukor/eP-ALM/utils.py
code served (permissive licence) · get_code("aa81d03f07735ab8")
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

torch.manual_seed(0)

# Create small tensors for testing
# logits: shape (batch_size=2, num_classes=4)
logits = torch.randn(2, 4)

# label: shape (batch_size=2,) with class indices
label = torch.tensor([1, 3])

# reduction: string parameter
reduction = 'mean'

ARGS = {
    'logits': logits,
    'label': label,
    'reduction': reduction
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_n_params Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/utils.py
code served (permissive licence) · get_code("ad12a494674d23fb")
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 str. 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)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    "model": model,
    "return_str": 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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
connector Ran
ran on a synthesized input
mshukor/eP-ALM/models/connector.py
code served (permissive licence) · get_code("e1cd32aef5c028ea")
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 ModuleList. 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
from torch import nn

torch.manual_seed(0)

ARGS = {
    'connector_type': 'linear',
    'input_dim': 4,
    'output_dim': 3,
    'num_layers': 2
}
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 ModuleList — a ModuleList output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ModuleList output the recorded ndim 1 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
filter_state Ran
ran on a synthesized input
mshukor/eP-ALM/models/utils.py
code served (permissive licence) · get_code("e3e2a8940b84fca0")
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 OrderedDict. 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 collections

torch.manual_seed(0)

# Create a sample state dict that mimics a typical model state_dict
# This will include keys with various prefixes to test the filtering logic
state = collections.OrderedDict([
    ('model_text.layer1.weight', torch.randn(4, 3)),
    ('model_text.layer1.bias', torch.randn(4)),
    ('transformer.encoder.weight', torch.randn(5, 6)),
    ('model_vision.conv.weight', torch.randn(7, 8)),
    ('adapter.linear.weight', torch.randn(3, 4)),
    ('adapter.linear.bias', torch.randn(3)),
    ('layer.output.weight', torch.randn(2, 2)),
    ('layer.output.bias', torch.randn(2)),
    ('other_model.weight', torch.randn(4, 5)),
])

exclude_list = ['model_text', 'transformer', 'model_vision']

ARGS = {
    'state': state,
    'exclude_list': exclude_list,
}
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 OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict 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_init_weights_vit Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/models/vit.py
pointer only (licence: MIT); the stored copy get_code serves, from skyofbeginning/fedmgp, is pointer only (licence: NONE) · get_code("6c600a2f0f2361f1")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from skyofbeginning/fedmgp, 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.

get_pairs Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/models/clip.py
code served (permissive licence) · get_code("d919ae32e5e4e616")
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 set. 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 function get_pairs takes a word parameter
# A word is represented as a tuple of symbols (variable-length strings)
# The function iterates through the word and creates pairs of consecutive symbols

word = ("hello", "world", "test")

ARGS = {
    "word": word
}
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 set — a set output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a set 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
print_trainable_params_percentage Ran
ran on a synthesized input
mshukor/eP-ALM/models/utils.py
code served (permissive licence) · get_code("6acafcd14371ead2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

# Freeze some parameters to test trainable vs total
model.linear1.weight.requires_grad = False

ARGS = {
    'model': model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
rank_answer Ran
ran on a synthesized input
mshukor/eP-ALM/models/epalm.py
code served (permissive licence) · get_code("95d5fda3063a7f2e")
How it ran

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

import torch
import torch.nn as nn
from transformers.tokenization_utils_base import BatchEncoding

torch.manual_seed(0)

class MockTokenizer:
    def __init__(self):
        self.pad_token_id = 0

class MockModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 100)
    
    def forward(self, image, inputs, return_dict=True, mode='evaluate', labels=None, reduction='none'):
        batch_size = inputs.input_ids.size(0)
        seq_len = inputs.input_ids.size(1)
        
        logits = self.linear(torch.randn(batch_size, seq_len, 4))
        logits = logits[:, :, :100]
        
        class Output:
            def __init__(self, logits_val, loss_val=None):
                self.logits = logits_val
                self.loss = loss_val
        
        if mode == 'evaluate':
            return Output(logits)
        elif mode == 'train':
            loss = torch.abs(logits.mean(dim=2)) * 0.1
            return Output(logits, loss)
        
        return Output(logits)

batch_size = 2
seq_len = 5
vocab_size = 100
num_answers = 3
answer_seq_len = 4
k = 2

model = MockModel()
image = torch.randn(batch_size, 4)
question_input = BatchEncoding({
    'input_ids': torch.randint(1, vocab_size, (batch_size, seq_len)),
    'attention_mask': torch.ones(batch_size, seq_len, dtype=torch.long)
})
answer_ids = torch.randint(1, vocab_size, (num_answers, answer_seq_len))
answer_atts = torch.ones(num_answers, answer_seq_len, dtype=torch.long)
tokenizer = MockTokenizer()

ARGS = {
    'model': model,
    'image': image,
    'question_input': question_input,
    'answer_ids': answer_ids,
    'answer_atts': answer_atts,
    'k': k,
    'tokenizer': tokenizer,
    'special_answer_token': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, transformers — 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
resize_pos_embed Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mshukor/eP-ALM/models/vit.py
code served (permissive licence) · get_code("321424d80e3f6eed")
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 3, dtype family float, shape (1, 17, 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
import math

torch.manual_seed(0)

# Small position embedding tensor with batch=1, seq_len=10, embedding_dim=4
# Structure: [batch, seq_len, embed_dim]
# With num_prefix_tokens=1, we have 1 prefix token + 9 grid tokens
# 9 = 3x3 grid, so sqrt(9) = 3
posemb = torch.randn(1, 10, 4)

# New position embedding tensor with different sequence length
# Let's say we want a 4x4 grid + 1 prefix token = 17 tokens
posemb_new = torch.randn(1, 17, 4)

# Number of prefix tokens (e.g., class token in ViT)
num_prefix_tokens = 1

# New grid size (height, width) for interpolation
# For 16 grid tokens: 4x4
gs_new = (4, 4)

ARGS = {
    'posemb': posemb,
    'posemb_new': posemb_new,
    'num_prefix_tokens': num_prefix_tokens,
    'gs_new': gs_new,
}
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 3, dtype family float, shape (1, 17, 4) — shape (1, 17, 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
tile Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/models/epalm.py
pointer only (licence: MIT); the stored copy get_code serves, from timmy-chan/miss, is pointer only (licence: NONE) · get_code("aa5a91226269bf75")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from timmy-chan/miss, 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.

tile Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/models/opt.py
code served (permissive licence) · get_code("41a81099fded79c2")
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 4, dtype family float, shape (2, 9, 4, 5) 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

torch.manual_seed(0)

# Create a small tensor to test the tile function
# Using distinct dimensions: batch=2, channels=3, height=4, width=5
x = torch.randn(2, 3, 4, 5)

# Parameters for the tile function:
# - x: the input tensor
# - dim: the dimension along which to tile (must be valid for x)
# - n_tile: number of times to tile

ARGS = {
    "x": x,
    "dim": 1,
    "n_tile": 3
}
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 Tensor, ndim 4, dtype family float, shape (2, 9, 4, 5) — shape (2, 9, 4, 5) 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
whitespace_clean Ran
ran; our contract draft was wrong, not the code
mshukor/eP-ALM/models/clip.py
pointer only (licence: MIT); the stored copy get_code serves, from ltpwy/MSCI, is pointer only (licence: NONE) · get_code("9542161e9640b858")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ltpwy/MSCI, 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.

filter_msg Not yet run
unverified (harvested, no recorded run)
mshukor/eP-ALM/models/utils.py
code served (permissive licence) · get_code("2b891d6e13741183")
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 15 distinct functions with an identity, 10 are also present in other papers whose code we hold (8 with at least one paper dated earlier, 6 the same month, 10 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
get_pairs 258 other papers
46 dated earlier, 4 the same month, 200 later, 7 same year, order not determinable, 1 undated · 4 only as an assembly brick
1802.06474 (2018-02)
A Closed-form Solution to Photorealistic Image Stylization
1803.01128 (2018-03)
Seq2Sick: Evaluating the Robustness of Sequence-to-Sequence Models with Adversarial Examples
1809.00125 (2018-09)
Simple Fusion: Return of the Language Model
and 255 more not named here
whitespace_clean 231 other papers
31 dated earlier, 4 the same month, 190 later, 5 same year, order not determinable, 1 undated · 4 only as an assembly brick
1901.07441 (2019-01)
PadChest: A large chest x-ray image dataset with multi-label annotated reports
2107.06383 (2021-07)
How Much Can CLIP Benefit Vision-and-Language Tasks?
2108.10904 (2021-08)
SimVLM: Simple Visual Language Model Pretraining with Weak Supervision
and 228 more not named here
basic_clean 218 other papers
31 dated earlier, 4 the same month, 177 later, 5 same year, order not determinable, 1 undated · 4 only as an assembly brick
1901.07441 (2019-01)
PadChest: A large chest x-ray image dataset with multi-label annotated reports
2107.06383 (2021-07)
How Much Can CLIP Benefit Vision-and-Language Tasks?
2108.10904 (2021-08)
SimVLM: Simple Visual Language Model Pretraining with Weak Supervision
and 215 more not named here
tile 84 other papers
6 dated earlier, 78 later
2206.07699 (2022-06)
Write and Paint: Generative Vision-Language Models are Unified Modal Learners
2210.07795 (2022-10)
EfficientVLM: Fast and Accurate Vision-Language Models via Knowledge Distillation and Modal-adaptive Pruning
2210.08773 (2022-10)
Plug-and-Play VQA: Zero-shot VQA by Conjoining Large Pretrained Models with Zero Training
and 81 more not named here
compute_n_params 29 other papers
7 dated earlier, 1 the same month, 21 later
2111.09452 (2021-11)
Open Vocabulary Object Detection with Pseudo Bounding-Box Labels
2211.09790 (2022-11)
ConStruct-VL: Data-Free Continual Structured VL Concepts Learning
2211.13594 (2022-11)
Self-supervised vision-language pretraining for Medical visual question answering
and 26 more not named here
compute_acc 28 other papers
7 dated earlier, 1 the same month, 20 later
2111.09452 (2021-11)
Open Vocabulary Object Detection with Pseudo Bounding-Box Labels
2211.09790 (2022-11)
ConStruct-VL: Data-Free Continual Structured VL Concepts Learning
2211.13594 (2022-11)
Self-supervised vision-language pretraining for Medical visual question answering
and 25 more not named here
get_init_weights_vit 25 other papers
3 dated earlier, 21 later, 1 same year, order not determinable · 1 only as an assembly brick
2112.08654 (2021-12)
Learning to Prompt for Continual Learning
2207.12819 (2022-07)
S-Prompts Learning with Pre-trained Transformers: An Occam's Razor for Domain Incremental Learning
2210.08823 (2022-10)
Scaling & Shifting Your Features: A New Baseline for Efficient Model Tuning
and 22 more not named here
all_gather 16 other papers
2 dated earlier, 1 the same month, 10 later, 3 same year, order not determinable
2108.02833 (2021-08)
Elaborative Rehearsal for Zero-shot Action Recognition
136730727 (2022 (year only))
Self-Supervised Sparse Representation for Video Anomaly Detection
2023.findings-emnlp.644 (2023 (year only))
VIP5: Towards Multimodal Foundation Models for Recommendation
and 13 more not named here
resize_pos_embed 3 other papers
3 later
2308.03374 (2023-08)
Heterogeneous Forgetting Compensation for Class-Incremental Learning
2308.09610 (2023-08)
On the Effectiveness of LayerNorm Tuning for Continual Learning in Vision Transformers
2311.12048 (2023-11)
One Size Fits All for Semantic Shifts: Adaptive Prompt Tuning for Continual Learning
checkpoint_filter_fn 1 other paper
1 later
2308.03374 (2023-08)
Heterogeneous Forgetting Compensation for Class-Incremental Learning

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Large Language Models (LLMs) have so far impressed the world, with unprecedented capabilities that emerge in models at large scales. On the vision side, transformer models (i.e., ViT) are following the same trend, achieving the best performance on challenging benchmarks. With the abundance of such unimodal models, a natural question arises; do we need also to follow this trend to tackle multimodal tasks? In this work, we propose to rather direct effort to efficient adaptations of existing models, and propose to augment Language Models with perception. Existing approaches for adapting pretrained models for vision-language tasks still rely on several key components that hinder their efficiency. In particular, they still train a large number of parameters, rely on large multimodal pretraining, use encoders (e.g., CLIP) trained on huge image-text datasets, and add significant inference overhead. In addition, most of these approaches have focused on Zero-Shot and In Context Learning, with little to no effort on direct finetuning. We investigate the minimal computational effort needed to adapt unimodal models for multimodal tasks and propose a new challenging setup, alongside different approaches, that efficiently adapts unimodal pretrained models. We show that by freezing more than 99% of total parameters, training only one linear projection layer, and prepending only one trainable token, our approach (dubbed eP-ALM) significantly outperforms other baselines on VQA and Captioning across Image, Video, and Audio modalities, following the proposed setup. The code is available here: https://github.com/mshukor/eP-ALM.

For agents

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

get_harvested_code_for_paper("2303.11403")
get_code_for_paper("2303.11403")
have("2303.11403")

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

curl https://syntology.ai/api/ran/2303.11403.json

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

[![Syntology run record](https://syntology.ai/api/ran/2303.11403.svg)](https://syntology.ai/paper/2303.11403)

Connect an agent — have() is free.