SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2404.01740 · ICLR · 2024

Weakly-supervised Audio Separation via Bi-modal Semantic Similarity

Saeed Amizadeh, Kazuhito Koishida, Diana Marculescu, Tanvir Mahmud

arXiv · PDF · Open in the Atlas

Code that ran

12 of 18 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (7 of them constructed an object rather than computing a result; 2 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 2 failures of our instrument; 6 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/2404.01740.json · badge: /api/ran/2404.01740.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
microsoft/bimodalaudioseparation canonical (classifier label)
classifier basis: classifier_2025
8 ran · 4 no recorded run
microsoft/BiModalAudioSeparation canonical (classifier label)
classifier basis: census
4 ran · 2 no recorded run
FunctionStatusWhere it lives
BlockCA Ran
ran on a synthesized input · constructed an object; no result was computed
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("84919ee649260509")
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, BlockCA: 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
import torch.nn as nn

torch.manual_seed(0)

# Define the Mlp class (needed as default for mlp_layer parameter)
class Mlp(nn.Module):
    """ MLP as used in Vision Transformer, MLP-Mixer and related networks
    """

    def __init__(self, in_features, hidden_features=None, act_layer=nn.GELU):
        super().__init__()
        self.fc1 = nn.Linear(in_features, hidden_features)
        self.act = act_layer()
        self.fc2 = nn.Linear(hidden_features, in_features)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        x = self.fc2(x)
        return x

# Build arguments for BlockCA
ARGS = {
    'dim': 4,
    'cond_dim': 3,
    'num_heads': 2,
    'mlp_ratio': 2.0,
    'act_layer': nn.GELU,
    'norm_layer': nn.LayerNorm,
    'mlp_layer': Mlp,
}
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 BlockCA — 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 BlockCA 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
BlockCat Ran
ran on a synthesized input · constructed an object; no result was computed
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("b733542e13e4b8ce")
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, BlockCat: 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
import torch.nn as nn

torch.manual_seed(0)

# Define the BlockCat module
class BlockCat(nn.Module):
    def __init__(self, dim, cond_dim, num_heads=None):
        super().__init__()
        self.conv1 = nn.Conv2d(dim, dim, kernel_size=3, stride=1, padding=1)
        self.lin = nn.Linear(cond_dim, dim)
        self.conv2 = nn.Conv2d(dim * 2, dim, kernel_size=3, stride=1, padding=1)

    def forward(self, x, cond):
        res = x
        x = self.conv1(x)
        B, C, H, W = x.size()
        cond = self.lin(cond)
        _, NC = cond.size()
        cond = cond.unsqueeze(-1).unsqueeze(-1) * torch.ones((B, NC, H, W), device=x.device)
        x = torch.concat((x, cond), dim=1)
        x = self.conv2(x)
        x = x + res
        return x

# Create arguments for BlockCat
dim = 4
cond_dim = 3
num_heads = None

# Create sample tensors
batch_size = 2
height = 8
width = 8
x = torch.randn(batch_size, dim, height, width)
cond = torch.randn(batch_size, cond_dim)

ARGS = {
    "dim": dim,
    "cond_dim": cond_dim,
    "num_heads": num_heads,
}
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 BlockCat — 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 BlockCat 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
BlockIdentity Ran
ran on a synthesized input · constructed an object; no result was computed
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("8ec43e0a25ffe65f")
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, BlockIdentity: 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
import torch.nn as nn

torch.manual_seed(0)

# Construct arguments for BlockIdentity
dim = 4
cond_dim = 3
num_heads = 2

ARGS = {
    "dim": dim,
    "cond_dim": cond_dim,
    "num_heads": num_heads,
}
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 BlockIdentity — 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 BlockIdentity 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
BlockSA Ran
ran on a synthesized input · constructed an object; no result was computed
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("903f2610f1453cea")
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, BlockSA: 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
import torch.nn as nn

torch.manual_seed(0)

# Define the MLP class as provided
class Mlp(nn.Module):
    def __init__(self, in_features, hidden_features=None, act_layer=nn.GELU):
        super().__init__()
        self.fc1 = nn.Linear(in_features, hidden_features)
        self.act = act_layer()
        self.fc2 = nn.Linear(hidden_features, in_features)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        x = self.fc2(x)
        return x

# Define the SelfAttention class as provided
class SelfAttention(nn.Module):
    def __init__(self, dim, num_heads=8):
        super().__init__()
        assert dim % num_heads == 0, 'dim should be divisible by num_heads'
        self.num_heads = num_heads
        self.head_dim = dim // num_heads
        self.scale = self.head_dim ** (-0.5)
        self.qkv = nn.Linear(dim, dim * 3)
        self.proj = nn.Linear(dim, dim)

    def forward(self, x):
        B, N, C = x.shape
        qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, self.head_dim).permute(2, 0, 3, 1, 4)
        q, k, v = qkv.unbind(0)
        q = q * self.scale
        attn = q @ k.transpose(-2, -1)
        attn = attn.softmax(dim=-1)
        x = attn @ v
        x = x.transpose(1, 2).reshape(B, N, C)
        x = self.proj(x)
        return x

# Define the BlockSA class as provided
class BlockSA(nn.Module):
    def __init__(self, dim, cond_dim, num_heads, mlp_ratio=2.0, act_layer=nn.GELU, norm_layer=nn.LayerNorm, mlp_layer=Mlp):
        super().__init__()
        self.norm1 = norm_layer(dim)
        self.attn = SelfAttention(dim, num_heads=num_heads)
        self.norm2 = norm_layer(dim)
        self.mlp = mlp_layer(in_features=dim, hidden_features=int(dim * mlp_ratio), act_layer=act_layer)

    def forward(self, x, cond=None):
        B, C, H, W = x.shape
        x = x.view(B, C, -1).permute(0, 2, 1)
        x = x + self.attn(self.norm1(x))
        x = x + self.mlp(self.norm2(x))
        x = x.permute(0, 2, 1).view(B, C, H, W)
        return x

# Construct ARGS for BlockSA
ARGS = {
    'dim': 4,
    'cond_dim': 8,
    'num_heads': 2,
    'mlp_ratio': 2.0,
    'act_layer': nn.GELU,
    'norm_layer': nn.LayerNorm,
    'mlp_layer': Mlp
}
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 BlockSA — 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 BlockSA 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
CrossAttention Ran
ran on a synthesized input · constructed an object; no result was computed
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("837dd2dcaeaaded4")
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, CrossAttention: 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
import torch.nn as nn

torch.manual_seed(0)

class CrossAttention(nn.Module):

    def __init__(self, dim, cond_dim, num_heads=8):
        super().__init__()
        assert dim % num_heads == 0, 'dim should be divisible by num_heads'
        self.num_heads = num_heads
        self.head_dim = dim // num_heads
        self.scale = self.head_dim ** (-0.5)
        self.q = nn.Linear(dim, dim)
        self.kv = nn.Linear(cond_dim, dim * 2)
        self.proj = nn.Linear(dim, dim)

    def forward(self, x, cond):
        _, condN, _ = cond.shape
        B, N, C = x.shape
        q = self.q(x).reshape(B, N, self.num_heads, self.head_dim).permute(0, 2, 1, 3)
        kv = self.kv(cond).reshape(B, condN, 2, self.num_heads, self.head_dim).permute(2, 0, 3, 1, 4)
        k, v = kv.unbind(0)
        q = q * self.scale
        attn = q @ k.transpose(-2, -1)
        attn = attn.softmax(dim=-1)
        x = attn @ v
        x = x.transpose(1, 2).reshape(B, N, C)
        x = self.proj(x)
        return x

dim = 8
cond_dim = 12
num_heads = 2

cross_attn = CrossAttention(dim=dim, cond_dim=cond_dim, num_heads=num_heads)

x = torch.randn(2, 5, 8)
cond = torch.randn(2, 7, 12)

ARGS = {
    'dim': dim,
    'cond_dim': cond_dim,
    'num_heads': num_heads,
}
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 CrossAttention — 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 CrossAttention 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
Mlp Ran
ran on a synthesized input · constructed an object; no result was computed
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("033bf6b0228555c7")
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, Mlp: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "in_features": 4,
    "hidden_features": 8,
    "act_layer": nn.GELU,
}
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 Mlp — 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 Mlp 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
ResBlock Ran
ran on a synthesized input · constructed an object; no result was computed
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("a84641b3af546ce0")
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, ResBlock: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "input_nc": 16,
    "output_nc": 16,
    "num_res_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 ResBlock — 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 ResBlock 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
SelfAttention Ran
ran on a synthesized input
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("fee297065b6d46a5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, 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

# Dimension for the model
dim = 4
num_heads = 2

# Arguments for __init__
INIT_ARGS = {
    'dim': dim,
    'num_heads': num_heads,
}

# Arguments for forward (the computation method)
# x should have shape (batch_size, sequence_length, dim)
ARGS = {
    'x': torch.randn(2, 3, dim),
}

# Construct and call the class
if __name__ == '__main__':
    model = nn.SelfAttention(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Input shape: {ARGS['x'].shape}")
    print(f"Output shape: {output.shape}")
    print("Success!")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
load_state_dict Ran
ran; our contract draft was wrong, not the code
microsoft/BiModalAudioSeparation/models/laion_clap/clap_module/factory.py
pointer only (licence: MIT); the stored copy get_code serves, from karchkha/msg-ld, is pointer only (licence: NONE) · get_code("af72484d5cfd7097")
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 karchkha/msg-ld, 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.

realimag Ran
ran, honoured the contract we drafted
microsoft/BiModalAudioSeparation/utils.py
code served (permissive licence) · get_code("81e9a172efc0d908")
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 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Construct small tensors for the realimag function
# mag: magnitude tensor
# cos: cosine values tensor
# sin: sine values tensor

mag = torch.randn(2, 4)
cos = torch.randn(2, 4)
sin = torch.randn(2, 4)

ARGS = {
    "mag": mag,
    "cos": cos,
    "sin": sin
}
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 3, 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
recover_rgb Ran
ran; our contract draft was wrong, not the code
microsoft/BiModalAudioSeparation/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from weiguoPian/ContAV-Sep_NeurIPS2024, is pointer only (licence: NONE) · get_code("3e5b3b6994917c24")
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 weiguoPian/ContAV-Sep_NeurIPS2024, 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.

warpgrid Ran
ran, honoured the contract we drafted
microsoft/BiModalAudioSeparation/utils.py
code served (permissive licence) · get_code("a4a08206de81190b")
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 ndarray, ndim 4, dtype family float, shape (2, 4, 3, 2) 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 numpy as np

# Construct arguments for warpgrid function
bs = 2  # batch size
HO = 4  # output height
WO = 3  # output width
warp = True  # apply warping transformation

ARGS = {
    "bs": bs,
    "HO": HO,
    "WO": WO,
    "warp": warp
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 4, dtype family float, shape (2, 4, 3, 2) — shape (2, 4, 3, 2) 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
BlockSCA Not yet run
unverified (harvested, no recorded run)
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("da16db8294ed4818")
How it ran

no invocation recorded.

BlockSCACat Not yet run
unverified (harvested, no recorded run)
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("e4e711f2a5fb5898")
How it ran

no invocation recorded.

CondAttUNet Not yet run
unverified (harvested, no recorded run)
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("9811209e675df9fc")
How it ran

no invocation recorded.

CondAttUNetBlock Not yet run
unverified (harvested, no recorded run)
microsoft/bimodalaudioseparation/models/cond_unet_attn.py
code served (permissive licence) · get_code("bdd91cc0612d6448")
How it ran

no invocation recorded.

create_model Not yet run
unverified (harvested, no recorded run)
microsoft/BiModalAudioSeparation/models/laion_clap/clap_module/factory.py
code served (permissive licence) · get_code("5c0daaa41ed490ea")
How it ran

no invocation recorded.

create_model_and_transforms Not yet run
unverified (harvested, no recorded run)
microsoft/BiModalAudioSeparation/models/laion_clap/clap_module/factory.py
code served (permissive licence) · get_code("ffdbb019791354d1")
How it ran

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, 5 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 5 later) and 13 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
warpgrid 6 other papers
4 dated earlier, 2 later · 2 only as an assembly brick
1908.11602 (2019-08)
Recursive Visual Sound Separation Using Minus-Plus Net
2007.07984 (2020-07)
Leveraging Category Information for Single-Frame Visual Sound Source Separation
2210.16472 (2022-10)
Learning Audio-Visual Dynamics Using Scene Graphs for Audio Source Separation
and 3 more not named here
create_model_and_transforms 4 other papers
1 dated earlier, 3 later
2308.01546 (2023-08)
MusicLDM: Enhancing Novelty in Text-to-Music Generation Using Beat-Synchronous Mixup Strategies
2405.18503 (2024-05)
SoundCTM: Unifying Score-based and Consistency Models for Full-band Text-to-Sound Generation
2409.12346 (2024-09)
Simultaneous Music Separation and Generation Using Multi-Track Latent Diffusion Models
and 1 more not named here
create_model 2 other papers
2 later
2405.18503 (2024-05)
SoundCTM: Unifying Score-based and Consistency Models for Full-band Text-to-Sound Generation
2605.30107 (2026-05)
Dial HEALTHDIAL for Advice: A Multilingual and Multi-Parallel Spoken Dialogue Dataset for Knowledge-Grounded Information Seeking
load_state_dict 2 other papers
1 dated earlier, 1 later
2308.01546 (2023-08)
MusicLDM: Enhancing Novelty in Text-to-Music Generation Using Beat-Synchronous Mixup Strategies
2409.12346 (2024-09)
Simultaneous Music Separation and Generation Using Multi-Track Latent Diffusion Models
recover_rgb 2 other papers
1 dated earlier, 1 later
1908.11602 (2019-08)
Recursive Visual Sound Separation Using Minus-Plus Net
2411.02860 (2024-11)
Continual Audio-Visual Sound Separation

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

Conditional sound separation in multi-source audio mixtures without having access to single source sound data during training is a long standing challenge. Existing mix-and-separate based methods suffer from significant performance drop with multi-source training mixtures due to the lack of supervision signal for single source separation cases during training. However, in the case of languageconditional audio separation, we do have access to corresponding text descriptions for each audio mixture in our training data, which can be seen as (rough) representations of the audio samples in the language modality. That raises the curious question of how to generate supervision signal for single-source audio extraction by leveraging the fact that single-source sounding language entities can be easily extracted from the text description. To this end, in this paper, we propose a generic bi-modal separation framework which can enhance the existing unsupervised frameworks to separate single-source signals in a target modality (i.e., audio) using the easily separable corresponding signals in the conditioning modality (i.e., language), without having access to single-source samples in the target modality during training. We empirically show that this is well within reach if we have access to a pretrained joint embedding model between the two modalities (i.e., CLAP). Furthermore, we propose to incorporate our framework into two fundamental scenarios to enhance separation performance. First, we show that our proposed methodology significantly improves the performance of purely unsupervised baselines by reducing the distribution shift between training and test samples. In particular, we show that our framework can achieve 71% boost in terms of Signal-to-Distortion Ratio (SDR) over the baseline, reaching 97.5% of the supervised learning performance. Second, we show that we can further improve the performance of the supervised learning itself by 17% if we augment it by our proposed weakly-supervised framework. Our framework achieves this by making large corpora of unsupervised data available to the supervised learning model as well as utilizing a natural, robust regularization mechanism through weak supervision from the language modality, and hence enabling a powerful semi-supervised framework for audio separation. Codes are available at https://github.com/microsoft/BiModalAudioSeparation/.

For agents

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

get_harvested_code_for_paper("2404.01740")
get_code_for_paper("2404.01740")
have("2404.01740")

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

curl https://syntology.ai/api/ran/2404.01740.json

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

[![Syntology run record](https://syntology.ai/api/ran/2404.01740.svg)](https://syntology.ai/paper/2404.01740)

Connect an agent — have() is free.