SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.09974 · 2024

GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery

arXiv · PDF · Open in the Atlas

Code that ran

7 of 14 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 0 have no contract verdict; 7 failures of our instrument; 7 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/2403.09974.json · badge: /api/ran/2403.09974.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
enguangw/get canonical (classifier label)
classifier basis: census
7 ran · 7 no recorded run
FunctionStatusWhere it lives
basic_clean Ran
ran; our contract draft was wrong, not the code
enguangw/get/clip/simple_tokenizer.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.

get_pairs Ran
ran; our contract draft was wrong, not the code
enguangw/get/clip/simple_tokenizer.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
get_params_groups Ran
ran; our contract draft was wrong, not the code
enguangw/get/model.py
code served (permissive licence) · get_code("ec22c16f1e0653f7")
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 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 torch.nn as nn

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        # Create parameters with different shapes to test the logic
        self.linear1 = nn.Linear(4, 3)  # weight: (3, 4), bias: (3,)
        self.linear2 = nn.Linear(3, 2)  # weight: (2, 3), bias: (2,)
        self.register_parameter('scalar_param', nn.Parameter(torch.tensor(1.0)))
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    'model': model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
info_nce_logits Ran
ran; our contract draft was wrong, not the code
enguangw/get/model.py
code served (permissive licence) · get_code("b3c2e94d4887d26b")
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. 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.functional as F

torch.manual_seed(0)

# The function expects:
# - features: a tensor of shape (batch_size, feature_dim)
#   The function does: features.size(0), F.normalize(features, dim=1), matmul with transpose
# - n_views: int, default 2
# - temperature: float, default 1.0
# - device: str, default 'cuda'

# Create features tensor
# With n_views=2, the function expects features.size(0) to be even (b_ = 0.5 * features.size(0))
# Let's use batch_size=4, feature_dim=8
features = torch.randn(4, 8)

ARGS = {
    'features': features,
    'n_views': 2,
    'temperature': 1.0,
    'device': 'cpu'
}
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
split_cluster_acc_v2 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
enguangw/get/util/cluster_and_log_utils.py
code served (permissive licence) · get_code("2513029b5bd17df9")
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 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 numpy as np

torch.manual_seed(0)

# Create sample data for the function
# y_true: ground truth labels
# y_pred: predicted labels
# mask: boolean mask indicating old vs new classes

# 10 samples total: 6 from old classes (mask=True), 4 from new classes (mask=False)
y_true = np.array([0, 0, 1, 1, 2, 2, 3, 3, 4, 4], dtype=np.int64)
y_pred = np.array([0, 0, 1, 1, 2, 2, 3, 3, 4, 4], dtype=np.int64)
mask = np.array([True, True, True, True, True, True, False, False, False, False], dtype=bool)

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "mask": mask,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, 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 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
split_cluster_acc_v2_balanced Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
enguangw/get/util/cluster_and_log_utils.py
code served (permissive licence) · get_code("a291ab390272cf0e")
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 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 numpy as np

torch.manual_seed(0)

# Create small sample data
n_samples = 10

# y_true: true labels with both old and new classes
y_true = np.array([0, 0, 1, 1, 2, 2, 3, 3, 4, 4], dtype=np.int64)

# y_pred: predicted labels (can have different mapping)
y_pred = np.array([0, 0, 1, 1, 2, 2, 3, 3, 4, 4], dtype=np.int64)

# mask: True for old classes (indices 0-5), False for new classes (indices 6-9)
mask = np.array([True, True, True, True, True, True, False, False, False, False], dtype=bool)

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "mask": mask,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, 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 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
whitespace_clean Ran
ran; our contract draft was wrong, not the code
enguangw/get/clip/simple_tokenizer.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.

all_sum_item Not yet run
unverified (harvested, no recorded run)
enguangw/get/util/cluster_and_log_utils.py
code served (permissive licence) · get_code("42be31a4153f5c2c")
How it ran

no invocation recorded.

build_TES_model Not yet run
unverified (harvested, no recorded run)
enguangw/get/clip/model.py
code served (permissive licence) · get_code("68683a9ee7224b56")
How it ran

no invocation recorded.

build_model Not yet run
unverified (harvested, no recorded run)
enguangw/get/clip/model.py
code served (permissive licence) · get_code("a059a5361ce98fb0")
How it ran

no invocation recorded.

distill_crit Not yet run
unverified (harvested, no recorded run)
enguangw/get/util/general_utils.py
code served (permissive licence) · get_code("5597baef301aa983")
How it ran

no invocation recorded.

init_experiment Not yet run
unverified (harvested, no recorded run)
enguangw/get/util/general_utils.py
code served (permissive licence) · get_code("39511f9d47e3ddc7")
How it ran

no invocation recorded.

load Not yet run
unverified (harvested, no recorded run)
enguangw/get/clip/clip.py
pointer only (licence: MIT); the stored copy get_code serves, from yusukekwn/Black-Box-Forgetting, is pointer only (licence: NONE) · get_code("2da7ec0975be872a")
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 yusukekwn/Black-Box-Forgetting, 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. No invocation recorded.

load_TES_CLIP Not yet run
unverified (harvested, no recorded run)
enguangw/get/clip/clip.py
code served (permissive licence) · get_code("bd0cff7a90b9726a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 14 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 5 the same month, 14 later) and 0 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
116 dated earlier, 14 the same month, 120 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
98 dated earlier, 13 the same month, 112 later, 7 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
98 dated earlier, 13 the same month, 99 later, 7 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
build_model 53 other papers
19 dated earlier, 32 later, 2 same year, order not determinable
2107.06383 (2021-07)
How Much Can CLIP Benefit Vision-and-Language Tasks?
2210.01115 (2022-10)
LASP: Text-to-Text Optimization for Language-Aware Soft Prompting of Vision & Language Models
2211.09552 (2022-11)
UniFormerV2: Spatiotemporal Learning by Arming Image ViTs with Video UniFormer
and 50 more not named here
load 18 other papers
4 dated earlier, 14 later
2107.06383 (2021-07)
How Much Can CLIP Benefit Vision-and-Language Tasks?
2311.11261 (2023-11)
Adversarial Prompt Tuning for Vision-Language Models
2311.18231 (2023-11)
TCP:Textual-based Class-aware Prompt tuning for Visual-Language Model
and 15 more not named here
get_params_groups 10 other papers
4 dated earlier, 1 the same month, 5 later · 1 only as an assembly brick
2104.14294 (2021-04)
Emerging Properties in Self-Supervised Vision Transformers
2211.11727 (2022-11)
Parametric Classification for Generalized Category Discovery: A Baseline Study
2308.12112 (2023-08)
Category Adaptation Meets Projected Distillation in Generalized Continual Category Discovery
and 7 more not named here
info_nce_logits 7 other papers
2 dated earlier, 1 the same month, 4 later · 1 only as an assembly brick
2211.11727 (2022-11)
Parametric Classification for Generalized Category Discovery: A Baseline Study
2310.01376 (2023-10)
Towards Distribution-Agnostic Generalized Category Discovery
2403.04272 (2024-03)
Active Generalized Category Discovery
and 4 more not named here
all_sum_item 1 other paper
1 later
Wang_GET_Unlocking_the_Multi-modal_Potential_of_CLIP_for_Generalized_Category_CVPR_2025_paper (2025 (year only))
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery
build_TES_model 1 other paper
1 later
Wang_GET_Unlocking_the_Multi-modal_Potential_of_CLIP_for_Generalized_Category_CVPR_2025_paper (2025 (year only))
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery
distill_crit 1 other paper
1 later
Wang_GET_Unlocking_the_Multi-modal_Potential_of_CLIP_for_Generalized_Category_CVPR_2025_paper (2025 (year only))
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery
init_experiment 1 other paper
1 later
Wang_GET_Unlocking_the_Multi-modal_Potential_of_CLIP_for_Generalized_Category_CVPR_2025_paper (2025 (year only))
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery
load_TES_CLIP 1 other paper
1 later
Wang_GET_Unlocking_the_Multi-modal_Potential_of_CLIP_for_Generalized_Category_CVPR_2025_paper (2025 (year only))
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery
split_cluster_acc_v2 1 other paper
1 later
Wang_GET_Unlocking_the_Multi-modal_Potential_of_CLIP_for_Generalized_Category_CVPR_2025_paper (2025 (year only))
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery
split_cluster_acc_v2_balanced 1 other paper
1 later
Wang_GET_Unlocking_the_Multi-modal_Potential_of_CLIP_for_Generalized_Category_CVPR_2025_paper (2025 (year only))
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery

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

Given unlabelled datasets containing both old and new categories, generalized category discovery (GCD) aims to accurately discover new classes while correctly classifying old classes. Current GCD methods only use a single visual modality of information, resulting in a poor classification of visually similar classes. As a different modality, text information can provide complementary discriminative information, which motivates us to introduce it into the GCD task. However, the lack of class names for unlabelled data makes it impractical to utilize text information. To tackle this challenging problem, in this paper, we propose a Text Embedding Synthesizer (TES) to generate pseudo text embeddings for unlabelled samples. Specifically, our TES leverages the property that CLIP can generate aligned vision-language features, converting visual embeddings into tokens of the CLIP's text encoder to generate pseudo text embeddings. Besides, we employ a dual-branch framework, through the joint learning and instance consistency of different modality branches, visual and semantic information mutually enhance each other, promoting the interaction and fusion of visual and text knowledge. Our method unlocks the multi-modal potentials of CLIP and outperforms the baseline methods by a large margin on all GCD benchmarks, achieving new state-of-the-art. Our code is available at: https://github.com/enguangW/GET.

For agents

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

get_harvested_code_for_paper("2403.09974")
get_code_for_paper("2403.09974")
have("2403.09974")

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

curl https://syntology.ai/api/ran/2403.09974.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.09974.svg)](https://syntology.ai/paper/2403.09974)

Connect an agent — have() is free.