Skip to content

added large, music and speech clap models #2284

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
isaac-chung merged 6 commits into from
Mar 8, 2025
Merged

added large, music and speech clap models #2284

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
cdde0c8
added large, music and speech clap models
RahulSChand Mar 8, 2025
a7302fe
fixed public_training_data and removed training_datasets split
RahulSChand Mar 8, 2025
736a25d
added latest revision
RahulSChand Mar 8, 2025
eeacb9d
lowercase mit license
RahulSChand Mar 8, 2025
2669e9e
fixed issue related to training_datasets
RahulSChand Mar 8, 2025
00c0464
fixed lint
RahulSChand Mar 8, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
98 changes: 88 additions & 10 deletions mteb/models/clap_models.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -194,44 +194,122 @@ def encode(
clap_htsat_fused = ModelMeta(
loader=partial(ClapZeroShotWrapper, model_name="laion/clap-htsat-fused"),
name="laion/clap-htsat-fused",
languages=["en"],
revision="main",
languages=["eng_Latn"],
revision="cca9e288ab447cee67d9ada1f85ddb46500f1401",
release_date="2023-05-22",
modalities=["audio", "text"],
n_parameters=153_507_530, # Calculated using torch.numel(model.parameters())
memory_usage_mb=586, # Calculated using model.calculate_memory_usage_mb()
max_tokens=float("inf"),
embed_dim=512, # The project_dim in config.json is 512
license="MIT",
license="mit",
open_weights=True,
public_training_code="https://github.com/LAION-AI/CLAP",
public_training_data="LAION-Audio-630K",
public_training_data="https://laion.ai/blog/laion-audio-630k/",
framework=["PyTorch"],
reference="https://huggingface.co/laion/clap_htsat_fused",
similarity_fn_name="cosine",
use_instructions=False,
training_datasets={"LAION-Audio-630K": ["https://laion.ai/blog/laion-audio-630k/"]},
training_datasets={
# "LAION-Audio-630K": ["https://laion.ai/blog/laion-audio-630k/"]
},
)


clap_htsat_unfused = ModelMeta(
loader=partial(ClapZeroShotWrapper, model_name="laion/clap-htsat-unfused"),
name="laion/clap-htsat-unfused",
languages=["en"],
revision="main",
languages=["eng_Latn"],
revision="8fa0f1c6d0433df6e97c127f64b2a1d6c0dcda8a",
release_date="2023-05-22",
modalities=["audio", "text"],
n_parameters=153_492_890, # Calculated using torch.numel(model.parameters())
memory_usage_mb=586, # Calculated using model.calculate_memory_usage_mb()
max_tokens=float("inf"),
embed_dim=512, # The project_dim in config.json is 512
license="MIT",
license="mit",
open_weights=True,
public_training_code="https://github.com/LAION-AI/CLAP",
public_training_data="LAION-Audio-630K",
public_training_data="https://laion.ai/blog/laion-audio-630k/",
framework=["PyTorch"],
reference="https://huggingface.co/laion/clap_htsat_unfused",
similarity_fn_name="cosine",
use_instructions=False,
training_datasets={"LAION-Audio-630K": ["https://laion.ai/blog/laion-audio-630k/"]},
training_datasets={
# "LAION-Audio-630K": ["https://laion.ai/blog/laion-audio-630k/"]
},
)

larger_clap_general = ModelMeta(
loader=partial(ClapZeroShotWrapper, model_name="laion/larger_clap_general"),
name="laion/larger_clap_general",
languages=["eng_Latn"],
revision="ada0c23a36c4e8582805bb38fec3905903f18b41",
release_date="2023-05-22",
modalities=["audio", "text"],
n_parameters=193_913_882, # Calculated using torch.numel(model.parameters())
memory_usage_mb=740, # Calculated using model.calculate_memory_usage_mb()
max_tokens=float("inf"),
embed_dim=512, # The project_dim (for even larger clap general) in config.json is 512
license="mit",
open_weights=True,
public_training_code="https://github.com/LAION-AI/CLAP",
public_training_data="https://laion.ai/blog/laion-audio-630k/",
framework=["PyTorch"],
reference="https://huggingface.co/laion/larger_clap_general",
similarity_fn_name="cosine",
use_instructions=False,
training_datasets={
# "LAION-Audio-630K": ["https://laion.ai/blog/laion-audio-630k/"]
}, # Additional finetuning over music dataset but not specified what the exact dataset is
)

larger_clap_music = ModelMeta(
loader=partial(ClapZeroShotWrapper, model_name="laion/larger_clap_music"),
name="laion/larger_clap_music",
languages=["eng_Latn"],
revision="a0b4534a14f58e20944452dff00a22a06ce629d1",
release_date="2023-05-22",
modalities=["audio", "text"],
n_parameters=193_913_882, # Calculated using torch.numel(model.parameters())
memory_usage_mb=740, # Calculated using model.calculate_memory_usage_mb()
max_tokens=float("inf"),
embed_dim=512, # The project_dim (for even larger clap general) in config.json is 512
license="mit",
open_weights=True,
public_training_code="https://github.com/LAION-AI/CLAP",
public_training_data="https://laion.ai/blog/laion-audio-630k/",
framework=["PyTorch"],
reference="https://huggingface.co/laion/larger_clap_music",
similarity_fn_name="cosine",
use_instructions=False,
training_datasets={
# "LAION-Audio-630K": ["https://laion.ai/blog/laion-audio-630k/"]
}, # Additional finetuning over music dataset but not specified what the exact dataset is
)

larger_clap_music_and_speech = ModelMeta(
loader=partial(
ClapZeroShotWrapper, model_name="laion/larger_clap_music_and_speech"
),
name="laion/larger_clap_music_and_speech",
languages=["eng_Latn"],
revision="195c3a3e68faebb3e2088b9a79e79b43ddbda76b",
release_date="2023-05-22",
modalities=["audio", "text"],
n_parameters=193_913_882, # Calculated using torch.numel(model.parameters())
memory_usage_mb=740, # Calculated using model.calculate_memory_usage_mb()
max_tokens=float("inf"),
embed_dim=512, # The project_dim (for even larger clap general) in config.json is 512
license="mit",
open_weights=True,
public_training_code="https://github.com/LAION-AI/CLAP",
public_training_data="https://laion.ai/blog/laion-audio-630k/",
framework=["PyTorch"],
reference="https://huggingface.co/laion/larger_clap_music_and_speech",
similarity_fn_name="cosine",
use_instructions=False,
training_datasets={
# "LAION-Audio-630K": ["https://laion.ai/blog/laion-audio-630k/"]
}, # Additional finetuning over music dataset but not specified what the exact dataset is
)
Loading