tomaarsen/clip-vit-L14-coco

CLIP ViT-L/14 model trained on COCO Captions

This is a sentence-transformers model finetuned from openai/clip-vit-large-patch14 on the coco_captions dataset. It maps sentences & paragraphs to a None-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: openai/clip-vit-large-patch14
• Maximum Sequence Length: None tokens
• Output Dimensionality: None dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • coco_captions
• Language: en
• License: apache-2.0

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'get_text_features', 'method_output_name': None}, 'image': {'method': 'get_image_features', 'method_output_name': None}}, 'module_output_name': 'sentence_embedding', 'architecture': 'CLIPModel'})
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("tomaarsen/clip-vit-L14-coco")
# Run inference
sentences = [
    'A large desk by a window is neatly arranged.',
    'A long hot dog on a plate on a table.',
    'A lady sitting at an enormous dining table with lots of food.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[ 1.0000, -0.0302,  0.1619],
#         [-0.0302,  1.0000,  0.1578],
#         [ 0.1619,  0.1578,  1.0000]])

Evaluation

Metrics

Information Retrieval

• Datasets: coco-eval and coco-test
• Evaluated with InformationRetrievalEvaluator

Metric	coco-eval	coco-test
cosine_accuracy@1	0.799	0.776
cosine_accuracy@3	0.968	0.959
cosine_accuracy@5	0.991	0.986
cosine_accuracy@10	0.995	0.995
cosine_precision@1	0.799	0.776
cosine_precision@3	0.3227	0.3197
cosine_precision@5	0.1982	0.1972
cosine_precision@10	0.0995	0.0995
cosine_recall@1	0.799	0.776
cosine_recall@3	0.968	0.959
cosine_recall@5	0.991	0.986
cosine_recall@10	0.995	0.995
cosine_ndcg@10	0.9112	0.8997
cosine_mrr@10	0.8827	0.8674
cosine_map@100	0.8828	0.8678

Training Details

Training Dataset

coco_captions

• Dataset: coco_captions at a2ed90d
• Size: 10,000 training samples
• Columns: image and caption
• Approximate statistics based on the first 1000 samples:

  	image	caption
  type	PIL.JpegImagePlugin.JpegImageFile	string
  details		min: 28 characters mean: 52.56 characters max: 156 characters

• Samples:

  image	caption
  	A woman wearing a net on her head cutting a cake.
  	A woman cutting a large white sheet cake.
  	A woman wearing a hair net cutting a large sheet cake.

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Evaluation Dataset

coco_captions

• Dataset: coco_captions at a2ed90d
• Size: 1,000 evaluation samples
• Columns: image and caption
• Approximate statistics based on the first 1000 samples:

  	image	caption
  type	PIL.JpegImagePlugin.JpegImageFile	string
  details		min: 27 characters mean: 52.45 characters max: 151 characters

• Samples:

  image	caption
  	A child holding a flowered umbrella and petting a yak.
  	A young man holding an umbrella next to a herd of cattle.
  	a young boy barefoot holding an umbrella touching the horn of a cow

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• learning_rate: 2e-05
• num_train_epochs: 1
• warmup_ratio: 0.1
• bf16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 2e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 1
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.1
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• use_cpu: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• bf16: True
• fp16: False
• half_precision_backend: None
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• optim_args: None
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• hub_revision: None
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_for_metrics: []
• eval_do_concat_batches: True
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: no
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• use_liger_kernel: False
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: True
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	coco-eval_cosine_ndcg@10	coco-test_cosine_ndcg@10
-1	-1	-	-	0.8902	-
0.0112	7	0.4782	-	-	-
0.0224	14	0.3108	-	-	-
0.0336	21	0.2212	-	-	-
0.0448	28	0.1612	-	-	-
0.056	35	0.1853	-	-	-
0.0672	42	0.0811	-	-	-
0.0784	49	0.0785	-	-	-
0.0896	56	0.1022	-	-	-
0.1008	63	0.0927	0.1433	0.9189	-
0.112	70	0.112	-	-	-
0.1232	77	0.1072	-	-	-
0.1344	84	0.1272	-	-	-
0.1456	91	0.1176	-	-	-
0.1568	98	0.1361	-	-	-
0.168	105	0.1281	-	-	-
0.1792	112	0.0961	-	-	-
0.1904	119	0.1038	-	-	-
0.2016	126	0.1019	0.1506	0.8929	-
0.2128	133	0.0657	-	-	-
0.224	140	0.1187	-	-	-
0.2352	147	0.0752	-	-	-
0.2464	154	0.2314	-	-	-
0.2576	161	0.0806	-	-	-
0.2688	168	0.1243	-	-	-
0.28	175	0.1179	-	-	-
0.2912	182	0.1174	-	-	-
0.3024	189	0.0926	0.1604	0.8907	-
0.3136	196	0.1327	-	-	-
0.3248	203	0.0861	-	-	-
0.336	210	0.0677	-	-	-
0.3472	217	0.1296	-	-	-
0.3584	224	0.1322	-	-	-
0.3696	231	0.1555	-	-	-
0.3808	238	0.0807	-	-	-
0.392	245	0.1134	-	-	-
0.4032	252	0.1826	0.1712	0.8840	-
0.4144	259	0.1796	-	-	-
0.4256	266	0.186	-	-	-
0.4368	273	0.0971	-	-	-
0.448	280	0.063	-	-	-
0.4592	287	0.1344	-	-	-
0.4704	294	0.072	-	-	-
0.4816	301	0.1233	-	-	-
0.4928	308	0.1152	-	-	-
0.504	315	0.148	0.1565	0.8960	-
0.5152	322	0.0836	-	-	-
0.5264	329	0.1171	-	-	-
0.5376	336	0.1433	-	-	-
0.5488	343	0.0494	-	-	-
0.56	350	0.1533	-	-	-
0.5712	357	0.0773	-	-	-
0.5824	364	0.0921	-	-	-
0.5936	371	0.0546	-	-	-
0.6048	378	0.1444	0.1496	0.9001	-
0.616	385	0.0956	-	-	-
0.6272	392	0.0445	-	-	-
0.6384	399	0.0939	-	-	-
0.6496	406	0.1109	-	-	-
0.6608	413	0.0466	-	-	-
0.672	420	0.0627	-	-	-
0.6832	427	0.0857	-	-	-
0.6944	434	0.058	-	-	-
0.7056	441	0.1542	0.1443	0.9031	-
0.7168	448	0.0972	-	-	-
0.728	455	0.0892	-	-	-
0.7392	462	0.0819	-	-	-
0.7504	469	0.0838	-	-	-
0.7616	476	0.0754	-	-	-
0.7728	483	0.0754	-	-	-
0.784	490	0.0638	-	-	-
0.7952	497	0.1006	-	-	-
0.8064	504	0.0398	0.1429	0.9122	-
0.8176	511	0.1562	-	-	-
0.8288	518	0.1039	-	-	-
0.84	525	0.0342	-	-	-
0.8512	532	0.0467	-	-	-
0.8624	539	0.0703	-	-	-
0.8736	546	0.0655	-	-	-
0.8848	553	0.0216	-	-	-
0.896	560	0.029	-	-	-
0.9072	567	0.0588	0.1530	0.9112	-
0.9184	574	0.1145	-	-	-
0.9296	581	0.0652	-	-	-
0.9408	588	0.0556	-	-	-
0.952	595	0.0458	-	-	-
0.9632	602	0.0085	-	-	-
0.9744	609	0.0572	-	-	-
0.9856	616	0.0942	-	-	-
0.9968	623	0.109	-	-	-
-1	-1	-	-	-	0.8997

Environmental Impact

Carbon emissions were measured using CodeCarbon.

• Energy Consumed: 0.043 kWh
• Carbon Emitted: 0.012 kg of CO2
• Hours Used: 0.137 hours

Training Hardware

• On Cloud: No
• GPU Model: 1 x NVIDIA GeForce RTX 3090
• CPU Model: 13th Gen Intel(R) Core(TM) i7-13700K
• RAM Size: 31.78 GB

Framework Versions

• Python: 3.11.6
• Sentence Transformers: 5.2.0.dev0
• Transformers: 4.57.0.dev0
• PyTorch: 2.8.0+cu128
• Accelerate: 1.6.0
• Datasets: 3.6.0
• Tokenizers: 0.22.1

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}