Skip to content

Commit

Permalink
[mieb] Fill in clip and open clip model meta (#1876)
Browse files Browse the repository at this point in the history 
* add clip and open clip model meta

* fix training_datasets
  • Loading branch information
@isaac-chung
isaac-chung authored Jan 26, 2025
1 parent b9fe9f0 commit 6ca11d2
Show file tree
Hide file tree
Showing 2 changed files with 114 additions and 98 deletions.
36 changes: 18 additions & 18 deletions mteb/models/clip_models.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -147,17 +147,17 @@ def get_fused_embeddings(
revision="32bd64288804d66eefd0ccbe215aa642df71cc41",
release_date="2021-02-26",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
n_parameters=428_000_000,
max_tokens=77,
embed_dim=768,
license=None,
open_weights=None,
open_weights=True,
public_training_code=None,
public_training_data=None,
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/openai/clip-vit-large-patch14",
similarity_fn_name=None,
use_instructions=None,
use_instructions=False,
training_datasets=None,
)

Expand All @@ -171,17 +171,17 @@ def get_fused_embeddings(
revision="3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268",
release_date="2021-02-26",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
n_parameters=151_000_000,
max_tokens=77,
embed_dim=512,
license=None,
open_weights=None,
open_weights=True,
public_training_code=None,
public_training_data=None,
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/openai/clip-vit-base-patch32",
similarity_fn_name=None,
use_instructions=None,
use_instructions=False,
training_datasets=None,
)

Expand All @@ -195,17 +195,17 @@ def get_fused_embeddings(
revision="57c216476eefef5ab752ec549e440a49ae4ae5f3",
release_date="2021-02-26",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
n_parameters=151_000_000,
max_tokens=77,
embed_dim=512,
license=None,
open_weights=None,
open_weights=True,
public_training_code=None,
public_training_data=None,
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/openai/clip-vit-base-patch16",
similarity_fn_name=None,
use_instructions=None,
use_instructions=False,
training_datasets=None,
)

Expand Down
176 changes: 96 additions & 80 deletions mteb/models/openclip_models.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -160,18 +160,20 @@ def get_fused_embeddings(
revision="84c9828e63dc9a9351d1fe637c346d4c1c4db341",
release_date="2023-04-26",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=428_000_000,
max_tokens=77,
embed_dim=768,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://huggingface.co/datasets/mlfoundations/datacomp_1b",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# DataComp-1B
},
)

CLIP_ViT_B_32_DataComp_XL_s13B_b90K = ModelMeta(
Expand All @@ -184,18 +186,20 @@ def get_fused_embeddings(
revision="f0e2ffa09cbadab3db6a261ec1ec56407ce42912",
release_date="2023-04-26",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=151_000_000,
max_tokens=77,
embed_dim=512,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://huggingface.co/datasets/mlfoundations/datacomp_1b",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# DataComp-1B
},
)

CLIP_ViT_B_16_DataComp_XL_s13B_b90K = ModelMeta(
Expand All @@ -208,18 +212,20 @@ def get_fused_embeddings(
revision="d110532e8d4ff91c574ee60a342323f28468b287",
release_date="2023-04-26",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=150_000_000,
max_tokens=77,
embed_dim=512,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://huggingface.co/datasets/mlfoundations/datacomp_1b",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-B-16-DataComp.XL-s13B-b90K",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# DataComp-1B
},
)

CLIP_ViT_bigG_14_laion2B_39B_b160k = ModelMeta(
Expand All @@ -232,18 +238,20 @@ def get_fused_embeddings(
revision="bc7788f151930d91b58474715fdce5524ad9a189",
release_date="2023-01-23",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=2_540_000_000,
max_tokens=77,
embed_dim=1280,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://laion.ai/blog/laion-5b/",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-bigG-14-laion2B-39B-b160k",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# 2 Billion sample English subset of LAION-5B
},
)

CLIP_ViT_g_14_laion2B_s34B_b88K = ModelMeta(
Expand All @@ -256,18 +264,20 @@ def get_fused_embeddings(
revision="15efd0f6ac0c40c0f9da7becca03c974d7012604",
release_date="2023-03-06",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=1_367_000_000,
max_tokens=77,
embed_dim=1024,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://laion.ai/blog/laion-5b/",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-g-14-laion2B-s34B-b88K",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# 2 Billion sample English subset of LAION-5B
},
)

CLIP_ViT_H_14_laion2B_s32B_b79K = ModelMeta(
Expand All @@ -280,18 +290,20 @@ def get_fused_embeddings(
revision="de081ac0a0ca8dc9d1533eed1ae884bb8ae1404b",
release_date="2022-09-15",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=986_000_000,
max_tokens=77,
embed_dim=1024,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://laion.ai/blog/laion-5b/",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# 2 Billion sample English subset of LAION-5B
},
)

CLIP_ViT_L_14_laion2B_s32B_b82K = ModelMeta(
Expand All @@ -304,18 +316,20 @@ def get_fused_embeddings(
revision="1627032197142fbe2a7cfec626f4ced3ae60d07a",
release_date="2022-09-15",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=428_000_000,
max_tokens=77,
embed_dim=768,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://laion.ai/blog/laion-5b/",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-L-14-laion2B-s32B-b82K",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# 2 Billion sample English subset of LAION-5B
},
)

CLIP_ViT_B_32_laion2B_s34B_b79K = ModelMeta(
Expand All @@ -328,16 +342,18 @@ def get_fused_embeddings(
revision="08f73555f1b2fb7c82058aebbd492887a94968ef",
release_date="2022-09-15",
modalities=["image", "text"],
n_parameters=None,
max_tokens=None,
embed_dim=None,
license=None,
open_weights=None,
public_training_code=None,
public_training_data=None,
n_parameters=151_000_000,
max_tokens=77,
embed_dim=512,
license="mit",
open_weights=True,
public_training_code="https://github.com/mlfoundations/open_clip",
public_training_data="https://laion.ai/blog/laion-5b/",
framework=["PyTorch"],
reference=None,
reference="https://huggingface.co/laion/CLIP-ViT-B-32-laion2B-s34B-b79K",
similarity_fn_name=None,
use_instructions=None,
training_datasets=None,
use_instructions=False,
training_datasets={
# 2 Billion sample English subset of LAION-5B
},
)

0 comments on commit 6ca11d2

Please sign in to comment.