SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.12656 · ACL · 2024

HyperMoE: Towards Better Mixture of Experts via Transferring Among Experts

Jie Fu, Hao Zhao, Zili Wang, Zhaofeng He, Zihan Qiu, Huijia Wu

arXiv · PDF · Open in the Atlas

Code that ran

14 of 18 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 11 failures of our instrument; 4 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2402.12656.json · badge: /api/ran/2402.12656.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
bumble666/hypermoe_early_version canonical (classifier label)
classifier basis not recorded on the link
11 ran · 3 no recorded run
Bumble666/Hyper_MoE community
Papers with Code archive
3 ran · 1 no recorded run
FunctionStatusWhere it lives
AdapterLayer Ran
ran on a synthesized input · constructed an object; no result was computed
Bumble666/Hyper_MoE/gpt-2-moe/transformers/models/layers.py
code served (permissive licence) · get_code("7fa2bef8aa39823f")
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 an instance of its own class, AdapterLayer: the run constructed the object and did not call it, so no result was computed. 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)

class Config:
    def __init__(self, adapter_dim):
        self.adapter_dim = adapter_dim

config = Config(adapter_dim=3)
input_size = 4
output_size = 5

ARGS = {
    "config": config,
    "input_size": input_size,
    "output_size": output_size,
}
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 AdapterLayer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a AdapterLayer 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
SparseDispatcher Ran
ran on a synthesized input · constructed an object; no result was computed
Bumble666/Hyper_MoE/gpt-2-moe/transformers/models/layers.py
code served (permissive licence) · get_code("3e2fbf364e529352")
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 an instance of its own class, SparseDispatcher: the run constructed the object and did not call it, so no result was computed. 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)

# Construct a gates tensor for the SparseDispatcher
# gates should be a 2D tensor where each row is a batch item and each column is an expert
# The dispatcher expects gates to have nonzero values indicating which experts are active
batch_size = 2
n_experts = 3

# Create a sparse gates matrix where some entries are nonzero
gates = torch.zeros(batch_size, n_experts)
gates[0, 0] = 0.5  # batch 0 uses expert 0
gates[0, 2] = 0.3  # batch 0 uses expert 2
gates[1, 1] = 0.7  # batch 1 uses expert 1
gates[1, 2] = 0.2  # batch 1 uses expert 2

ARGS = {
    "n_experts": n_experts,
    "gates": gates,
}
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 SparseDispatcher — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SparseDispatcher 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
booleans_processing Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_tf_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from clin0212/hydralora, is pointer only (licence: NONE) · get_code("743ceb9025a7955d")
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 clin0212/hydralora, 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.

convert_pytorch_state_dict_to_flax Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_flax_pytorch_utils.py
code served (permissive licence) · get_code("5168a172bbb363a5")
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
import jax.numpy as jnp
import numpy as np
from flax.traverse_util import flatten_dict, unflatten_dict

torch.manual_seed(0)

# Create a minimal PyTorch state dict
pt_state_dict = {
    'layer.0.weight': torch.randn(4, 3),
    'layer.0.bias': torch.randn(4),
    'layer.1.weight': torch.randn(5, 4),
    'layer.1.bias': torch.randn(5),
}

# Create a minimal Flax model class
class MinimalFlaxModel:
    def __init__(self):
        self.base_model_prefix = 'layer'
        
        # Build minimal params structure
        params_dict = {
            ('layer', '0', 'kernel'): jnp.ones((3, 4)),
            ('layer', '0', 'bias'): jnp.ones((4,)),
            ('layer', '1', 'kernel'): jnp.ones((4, 5)),
            ('layer', '1', 'bias'): jnp.ones((5,)),
        }
        
        # Unflatten to get nested structure
        unflat = unflatten_dict(params_dict)
        self.params = {'params': unflat}

flax_model = MinimalFlaxModel()

ARGS = {
    'pt_state_dict': pt_state_dict,
    'flax_model': flax_model,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
flax, jax, numpy, 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
dummy_loss Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_tf_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from clin0212/hydralora, is pointer only (licence: NONE) · get_code("3c1eb64dde6b86af")
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 clin0212/hydralora, 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_imports Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/dynamic_module_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from clin0212/hydralora, is pointer only (licence: NONE) · get_code("ccaa0edba373622c")
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 clin0212/hydralora, 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_relative_import_files Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/dynamic_module_utils.py
code served (permissive licence) · get_code("00cc853a73f3aeb6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re
from pathlib import Path
import tempfile
import os

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

# Create a simple module file with relative imports
module_file = os.path.join(temp_dir, "main_module.py")
with open(module_file, 'w') as f:
    f.write("""
from .helper import some_function
import .utils
from .submodule import ClassA
""")

# Create helper module
helper_file = os.path.join(temp_dir, "helper.py")
with open(helper_file, 'w') as f:
    f.write("""
def some_function():
    pass
""")

# Create utils module
utils_file = os.path.join(temp_dir, "utils.py")
with open(utils_file, 'w') as f:
    f.write("""
# utility functions
""")

# Create submodule
submodule_file = os.path.join(temp_dir, "submodule.py")
with open(submodule_file, 'w') as f:
    f.write("""
class ClassA:
    pass
""")

ARGS = {
    "module_file": module_file
}
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 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
get_relative_imports Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/dynamic_module_utils.py
code served (permissive licence) · get_code("450aeffe7f80f917")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import os

# Create a temporary Python file with relative imports
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test_module.py")

content = """
import .submodule
import .utils
from .helpers import some_function
from .core import another_function
import os
from typing import List
"""

with open(temp_file, 'w', encoding='utf-8') as f:
    f.write(content)

ARGS = {
    'module_file': temp_file
}
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 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
infer_metric_tags_from_eval_results Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modelcard.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from shen8424/CORE, is pointer only (licence: NONE) · get_code("dd8f7751ef859581")
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 shen8424/CORE, 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.

keras_serializable Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_tf_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from clin0212/hydralora, is pointer only (licence: NONE) · get_code("6a3959a0ffd36577")
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 clin0212/hydralora, 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.

noisy_top_k_gating_mixing Ran
ran; our contract draft was wrong, not the code
Bumble666/Hyper_MoE/gpt-2-moe/transformers/models/layers.py
code served (permissive licence) · get_code("7e4c76ff93dbad95")
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
from torch import nn

torch.manual_seed(0)

class MockLayer(nn.Module):
    def __init__(self, input_dim, n_experts, k, noisy_gating=True):
        super().__init__()
        self.n_experts = n_experts
        self.k = k
        self.noisy_gating = noisy_gating
        self.w_gate = nn.Parameter(torch.randn(input_dim, n_experts))
        self.w_noise = nn.Parameter(torch.randn(input_dim, n_experts))
        self.mean = torch.tensor(0.0)
        self.std = torch.tensor(1.0)

batch_size = 2
input_dim = 4
n_experts = 6
k = 2

layer = MockLayer(input_dim, n_experts, k, noisy_gating=True)
x = torch.randn(batch_size, input_dim)
train = True

ARGS = {
    "layer": layer,
    "x": x,
    "train": train,
    "noise_epsilon": 0.01,
}
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
parse_keras_history Ran
ran; our contract draft was wrong, not the code
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modelcard.py
code served (permissive licence) · get_code("a4491b45d19b7a39")
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 unittest.mock import Mock

torch.manual_seed(0)

# Create a mock tf.keras.History object
mock_history = Mock()
mock_history.history = {
    'loss': [0.5, 0.4, 0.3],
    'accuracy': [0.7, 0.8, 0.85],
    'val_loss': [0.6, 0.5, 0.4],
    'val_accuracy': [0.65, 0.75, 0.82]
}
mock_history.epoch = [0, 1, 2]

ARGS = {
    'logs': mock_history
}
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
quick_gelu Ran
ran, honoured the contract we drafted
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_flax_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from shen8424/CORE, is pointer only (licence: NONE) · get_code("a66d2b78670c85e1")
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 shen8424/CORE, 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.

rename_key_and_reshape_tensor Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_flax_pytorch_utils.py
code served (permissive licence) · get_code("61779b0406a59cfc")
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 jax.numpy as jnp
import numpy as np

# Create a small random Flax state dict with various key patterns
random_flax_state_dict = {
    ('layer', 'scale'): jnp.array([1.0, 2.0, 3.0, 4.0]),
    ('layer', 'mean'): jnp.array([0.1, 0.2, 0.3, 0.4]),
    ('layer', 'var'): jnp.array([0.01, 0.02, 0.03, 0.04]),
    ('layer', 'embedding'): jnp.ones((4, 5)),
    ('layer', 'kernel'): jnp.ones((2, 3, 4, 5)),
    ('prefix', 'layer', 'scale'): jnp.array([1.0, 2.0]),
}

# Create a PyTorch tensor to be renamed and potentially reshaped
# Using a 4D tensor for testing the 4D reshape case
pt_tensor = np.random.randn(3, 4, 5, 6).astype(np.float32)

# Create a tuple key representing a PyTorch weight name
pt_tuple_key = ('layer', 'weight')

# Model prefix used for checking prefixed keys
model_prefix = 'prefix'

ARGS = {
    'pt_tuple_key': pt_tuple_key,
    'pt_tensor': pt_tensor,
    'random_flax_state_dict': random_flax_state_dict,
    'model_prefix': model_prefix,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — 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
MoE Not yet run
unverified (harvested, no recorded run)
Bumble666/Hyper_MoE/gpt-2-moe/transformers/models/layers.py
code served (permissive licence) · get_code("89f7e4ccbdd62673")
How it ran

no invocation recorded.

dtype_byte_size Not yet run
unverified (harvested, no recorded run)
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_flax_utils.py
code served (permissive licence) · get_code("9d98806acb3144b7")
How it ran

no invocation recorded.

flax_shard_checkpoint Not yet run
unverified (harvested, no recorded run)
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modeling_flax_utils.py
code served (permissive licence) · get_code("e3ce70ca9258800e")
How it ran

no invocation recorded.

is_hf_dataset Not yet run
unverified (harvested, no recorded run)
bumble666/hypermoe_early_version/gpt-2-moe/transformers/modelcard.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from shen8424/CORE, is pointer only (licence: NONE) · get_code("ec7b1528a936234d")
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 shen8424/CORE, 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.

Also present in other papers' code

Of 18 samples lifted from this paper's linked repositories, 18 carry a function identity and 0 do not. Of the 18 distinct functions with an identity, 14 are also present in other papers whose code we hold (14 with at least one paper dated earlier, 10 the same month, 14 later) and 4 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
infer_metric_tags_from_eval_results 19 other papers
6 dated earlier, 1 the same month, 12 later
2022.findings-emnlp.427 (2022 (year only))
2305.15756 (2023-05)
UniTRec: A Unified Text-to-Text Transformer and Joint Contrastive Learning Framework for Text-based Recommendation
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
and 16 more not named here
get_relative_import_files 18 other papers
9 dated earlier, 1 the same month, 8 later
2211.11694 (2022-11)
Exploring Discrete Diffusion Models for Image Captioning
2304.05568 (2023-04)
Improving Diffusion Models for Scene Text Editing with Dual Encoders
2305.15756 (2023-05)
UniTRec: A Unified Text-to-Text Transformer and Joint Contrastive Learning Framework for Text-based Recommendation
and 15 more not named here
get_relative_imports 18 other papers
9 dated earlier, 1 the same month, 8 later
2211.11694 (2022-11)
Exploring Discrete Diffusion Models for Image Captioning
2304.05568 (2023-04)
Improving Diffusion Models for Scene Text Editing with Dual Encoders
2305.15756 (2023-05)
UniTRec: A Unified Text-to-Text Transformer and Joint Contrastive Learning Framework for Text-based Recommendation
and 15 more not named here
quick_gelu 17 other papers
7 dated earlier, 1 the same month, 9 later
2022.findings-emnlp.427 (2022 (year only))
2303.14582 (2023-03)
Identification of Negative Transfers in Multitask Learning Using Surrogate Models
2305.15756 (2023-05)
UniTRec: A Unified Text-to-Text Transformer and Joint Contrastive Learning Framework for Text-based Recommendation
and 14 more not named here
is_hf_dataset 16 other papers
3 dated earlier, 1 the same month, 12 later
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
2312.09979 (2023-12)
LoRAMoE: Alleviate World Knowledge Forgetting in Large Language Models via MoE-Style Plugin
and 13 more not named here
parse_keras_history 16 other papers
3 dated earlier, 1 the same month, 12 later
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
2312.09979 (2023-12)
LoRAMoE: Alleviate World Knowledge Forgetting in Large Language Models via MoE-Style Plugin
and 13 more not named here
flax_shard_checkpoint 14 other papers
5 dated earlier, 1 the same month, 8 later
2305.15756 (2023-05)
UniTRec: A Unified Text-to-Text Transformer and Joint Contrastive Learning Framework for Text-based Recommendation
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
and 11 more not named here
dtype_byte_size 12 other papers
5 dated earlier, 1 the same month, 6 later
2305.15756 (2023-05)
UniTRec: A Unified Text-to-Text Transformer and Joint Contrastive Learning Framework for Text-based Recommendation
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
and 9 more not named here
keras_serializable 8 other papers
5 dated earlier, 3 later
2109.03808 (2021-09)
Smelting Gold and Silver for Improved Multilingual AMR-to-Text Generation
2303.14582 (2023-03)
Identification of Negative Transfers in Multitask Learning Using Surrogate Models
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
and 5 more not named here
booleans_processing 6 other papers
3 dated earlier, 3 later
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
2310.18738 (2023-10)
TLM: Token-Level Masking for Transformers
and 3 more not named here
convert_pytorch_state_dict_to_flax 6 other papers
3 dated earlier, 1 the same month, 2 later
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
2312.09979 (2023-12)
LoRAMoE: Alleviate World Knowledge Forgetting in Large Language Models via MoE-Style Plugin
and 3 more not named here
dummy_loss 6 other papers
3 dated earlier, 3 later
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
2310.18738 (2023-10)
TLM: Token-Level Masking for Transformers
and 3 more not named here
rename_key_and_reshape_tensor 6 other papers
3 dated earlier, 1 the same month, 2 later
2310.05107 (2023-10)
OV-PARTS: Towards Open-Vocabulary Part Segmentation
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
2312.09979 (2023-12)
LoRAMoE: Alleviate World Knowledge Forgetting in Large Language Models via MoE-Style Plugin
and 3 more not named here
get_imports 3 other papers
2 dated earlier, 1 later
2310.09832 (2023-10)
Merging Experts into One: Improving Computational Efficiency of Mixture of Experts
2312.09979 (2023-12)
LoRAMoE: Alleviate World Knowledge Forgetting in Large Language Models via MoE-Style Plugin
2404.19245 (2024-04)
HydraLoRA: An Asymmetric LoRA Architecture for Efficient Fine-Tuning

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

The Mixture of Experts (MoE) for language models has been proven effective in augmenting the capacity of models by dynamically routing each input token to a specific subset of experts for processing. Despite the success, most existing methods face a challenge for balance between sparsity and the availability of expert knowledge: enhancing performance through increased use of expert knowledge often results in diminishing sparsity during expert selection. To mitigate this contradiction, we propose HyperMoE, a novel MoE framework built upon Hypernetworks. This framework integrates the computational processes of MoE with the concept of knowledge transferring in multi-task learning. Specific modules generated based on the information of unselected experts serve as supplementary information, which allows the knowledge of experts not selected to be used while maintaining selection sparsity. Our comprehensive empirical evaluations across multiple datasets and backbones establish that HyperMoE significantly outperforms existing MoE methods under identical conditions concerning the number of experts. Our code is publicly available at https: //github.com/Bumble666/Hyper_MoE

For agents

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

get_harvested_code_for_paper("2402.12656")
get_code_for_paper("2402.12656")
have("2402.12656")

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

curl https://syntology.ai/api/ran/2402.12656.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.12656.svg)](https://syntology.ai/paper/2402.12656)

Connect an agent — have() is free.