Table Continuation Classifier based on Ettin-32M

This is a Cross Encoder model finetuned from jhu-clsp/ettin-encoder-32m on the table-continuation-dataset dataset using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text reranking and semantic search.

Model Details

Model Description

Model Type: Cross Encoder
Base model: jhu-clsp/ettin-encoder-32m
Maximum Sequence Length: 4096 tokens
Number of Output Labels: 1 label
Training Dataset:
- table-continuation-dataset
Language: en
License: apache-2.0

Model Sources

Documentation: Sentence Transformers Documentation
Documentation: Cross Encoder Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Cross Encoders on Hugging Face

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 CrossEncoder

# Download from the 🤗 Hub
model = CrossEncoder("BioMike/table-linker-32m-init")
# Get scores for pairs of texts
pairs = [
    ['| K40Specs |  |\n| --- | --- |\n| GlobalMemory | 11520MB |\n| L2cachesize | 1.57MB |\n| Sharedmemoryperblock | 0.049MB |', '| Maxsharedmemoryperthreadblock | 49152bytes |\n| --- | --- |\n| L2cachesize | 1.50MiB |'],
    ['| method | mean() | median() | trimean() |\n| --- | --- | --- | --- |\n| Originallyreportedresults |  |  |  |\n| Shades-of-Gray | 11.55 | 9.70 | 10.23 |\n| GeneralGray-World | 11.55 | 9.70 | 10.23 |\n| 1st-orderGray-Edge | 10.58 | 8.84 | 9.18 |\n| 2nd-orderGray-Edge | 10.68 | 9.02 | 9.40 |\n| Revisitedresults |  |  |  |\n| Shades-of-Gray | 13.32 | 11.57 | 12.10 |\n| GeneralGray-World | 13.69 | 12.11 | 12.55 |\n| 1st-orderGray-Edge | 11.06 | 9.54 | 9.81 |\n| 2nd-orderGray-Edge | 10.73 | 9.21 | 9.49 |', '| Greenstabilityassumptionresults |  |  |  |\n| --- | --- | --- | --- |\n| Shades-of-Gray | 12.68 | 10.50 | 11.25 |\n| GeneralGray-World | 12.68 | 10.50 | 11.25 |\n| 1st-orderGray-Edge | 13.41 | 11.04 | 11.87 |\n| 2nd-orderGray-Edge | 12.83 | 10.70 | 11.44 |'],
    ['| Name | Abbreviation | #ofinstances |\n| --- | --- | --- |\n| wcsp/spot5/dir | wcsp-dir | 21 |\n| wcsp/spot5/log | wcsp-log | 21 |\n| haplotyping-pedigrees | HT | 100 |\n| upgradeability-problem | UP | 100 |\n| preferenceplanning | PP | 29 |', '| packup-wpms | PWPMS | 99 |\n| --- | --- | --- |\n| timetabling | TT | 26 |'],
    ['| Algorithm | NMSE(dB) | ComputationTime(sec) |\n| --- | --- | --- |\n| HMT+IRWL1 | -14.37 | 363 |\n| CoSaMP | -16.90 | 25 |\n| ModelCS | -17.45 | 117 |', '|  | Algorithm | Symm | Asymm |  |  |  |  |\n| --- | --- | --- | --- | --- | --- | --- | --- |\n| LCC | SROCC | RMSE | LCC | SROCC | RMSE |  |  |\n| VQUEMODES | SBIQE | 0.9000 | 0.8913 | 4.5900 | 0.8532 | 0.8234 | 7.1376 |\n| BRISQUE | 0.9125 | 0.9013 | 4.3980 | 0.8792 | 0.8489 | 6.8563 |  |\n| NIQE | 0.9285 | 0.9236 | 3.9852 | 0.8955 | 0.8490 | 6.9563 |  |'],
    ['| Country | PowerDist. | Individualism | Masculinity |\n| --- | --- | --- | --- |\n| Chile | 63 | 23 | 28 |\n| China | 80 | 20 | 66 |\n| Germany | 35 | 67 | 66 |\n| Greece | 60 | 35 | 57 |\n| India | 77 | 48 | 56 |', '| Country | dataset | dataset |  |  |\n| --- | --- | --- | --- | --- |\n| - | ρ | p-value | ρ | p-value |\n| Argentina | 0.56 | 0.03 | 0.77 | 0.0007 |\n| Australia | 0.32 | 0.23 | 0.60 | 0.02 |\n| Brazil | 0.48 | 0.06 | 0.81 | 0.0002 |\n| Chile | 0.32 | 0.23 | 0.53 | 0.04 |\n| England | 0.87 | 0 | 0.70 | 0.004 |\n| France | 0.85 | 2e-06 | 0.61 | 0.01 |\n| Indonesia | 0.84 | 4e-05 | 0.75 | 0.001 |\n| Japan | 0.38 | 0.15 | 0.39 | 0.13 |\n| Korea | 0.68 | 0.004 | 0.45 | 0.08 |\n| Malaysia | -0.16 | 0.54 | 0.11 | 0.68 |\n| Mexico | 0.55 | 0.03 | 0.71 | 0.003 |\n| Russia | 0.78 | 0.0006 | 0.76 | 0.001 |\n| Singapore | 0.34 | 0.20 | 0.65 | 0.008 |\n| Spain | 0.78 | 0.0005 | 0.75 | 0.001 |\n| Turkey | -0.18 | 0.50 | -0.31 | 0.24 |\n| USA | 0.70 | 0.004 | 0.67 | 0.005 |'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    '| K40Specs |  |\n| --- | --- |\n| GlobalMemory | 11520MB |\n| L2cachesize | 1.57MB |\n| Sharedmemoryperblock | 0.049MB |',
    [
        '| Maxsharedmemoryperthreadblock | 49152bytes |\n| --- | --- |\n| L2cachesize | 1.50MiB |',
        '| Greenstabilityassumptionresults |  |  |  |\n| --- | --- | --- | --- |\n| Shades-of-Gray | 12.68 | 10.50 | 11.25 |\n| GeneralGray-World | 12.68 | 10.50 | 11.25 |\n| 1st-orderGray-Edge | 13.41 | 11.04 | 11.87 |\n| 2nd-orderGray-Edge | 12.83 | 10.70 | 11.44 |',
        '| packup-wpms | PWPMS | 99 |\n| --- | --- | --- |\n| timetabling | TT | 26 |',
        '|  | Algorithm | Symm | Asymm |  |  |  |  |\n| --- | --- | --- | --- | --- | --- | --- | --- |\n| LCC | SROCC | RMSE | LCC | SROCC | RMSE |  |  |\n| VQUEMODES | SBIQE | 0.9000 | 0.8913 | 4.5900 | 0.8532 | 0.8234 | 7.1376 |\n| BRISQUE | 0.9125 | 0.9013 | 4.3980 | 0.8792 | 0.8489 | 6.8563 |  |\n| NIQE | 0.9285 | 0.9236 | 3.9852 | 0.8955 | 0.8490 | 6.9563 |  |',
        '| Country | dataset | dataset |  |  |\n| --- | --- | --- | --- | --- |\n| - | ρ | p-value | ρ | p-value |\n| Argentina | 0.56 | 0.03 | 0.77 | 0.0007 |\n| Australia | 0.32 | 0.23 | 0.60 | 0.02 |\n| Brazil | 0.48 | 0.06 | 0.81 | 0.0002 |\n| Chile | 0.32 | 0.23 | 0.53 | 0.04 |\n| England | 0.87 | 0 | 0.70 | 0.004 |\n| France | 0.85 | 2e-06 | 0.61 | 0.01 |\n| Indonesia | 0.84 | 4e-05 | 0.75 | 0.001 |\n| Japan | 0.38 | 0.15 | 0.39 | 0.13 |\n| Korea | 0.68 | 0.004 | 0.45 | 0.08 |\n| Malaysia | -0.16 | 0.54 | 0.11 | 0.68 |\n| Mexico | 0.55 | 0.03 | 0.71 | 0.003 |\n| Russia | 0.78 | 0.0006 | 0.76 | 0.001 |\n| Singapore | 0.34 | 0.20 | 0.65 | 0.008 |\n| Spain | 0.78 | 0.0005 | 0.75 | 0.001 |\n| Turkey | -0.18 | 0.50 | -0.31 | 0.24 |\n| USA | 0.70 | 0.004 | 0.67 | 0.005 |',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Evaluation

Metrics

Cross Encoder Binary Classification

Dataset: eval
Evaluated with CEBinaryClassificationEvaluator

Metric	Value
accuracy	0.8943
accuracy_threshold	0.4495
f1	0.9063
f1_threshold	0.116
precision	0.8387
recall	0.9857
average_precision	0.9266

Training Details

Training Dataset

table-continuation-dataset

Dataset: table-continuation-dataset at 45a5089
Size: 13,194 training samples
Columns: premise, hypothesis, and label

Approximate statistics based on the first 1000 samples:

	premise	hypothesis	label
type	string	string	int
details	min: 26 characters mean: 255.89 characters max: 3690 characters	min: 26 characters mean: 350.07 characters max: 2658 characters	0: ~49.70% 1: ~50.30%

Samples:

premise	hypothesis	label
`\| Networkarchitecture \| mAP \| \| --- \| --- \| \| StackedRNNwithelement-wisemax \| 58.1 \| \| RNNwithelement-wisemax \| 57.4 \|`	`\| Layer \| Time(seconds) \| \| --- \| --- \| \| Convlayer(32featuremaps) \| 29 \| \| Convlayer(32featuremaps) \| 755 \| \| ActivationLayer \| 113 \| \| AveragePoolingLayer \| 15 \| \| Convlayer(64featuremaps) \| 249 \| \| Convlayer(64featuremaps) \| 314 \| \| ActivationLayer \| 127 \| \| Convlayer(128featuremaps) \| 418 \| \| Convlayer(128featuremaps) \| 405 \| \| ActivationLayer \| 101 \| \| AveragePoolingLayer \| 3 \| \| 2FullyConnected(256and10neurons) \| 32 \|`	`0`
`\| Portion \| 1Annot. \| 2Annot. \| 3Annot. \| \| --- \| --- \| --- \| --- \| \| 1-2400 \| #08 \| #01 \| #04 \| \| 2401-4800 \| #04 \| #03 \| #01 \| \| 4801-7200 \| #01 \| #04 \| #08 \| \| 7201-9600 \| #03 \| #08 \| #02 \|`	`\| 623 \| 38 \| \| --- \| --- \| \| 405 \| 12 \| \| 527 \| 22 \| \| 491 \| 24 \| \| 531 \| 23 \| \| 474 \| 27 \| \| 515 \| 22 \| \| 856 \| 104 \| \| 424 \| 17 \| \| 515 \| 37 \| \| 403 \| 17 \| \| 1615 \| 186 \|`	`0`
`\| (b,b,b,b)0123 \| Typeoffunction \| Querycomplexity \| \| --- \| --- \| --- \| \| 0000 \| Constantfunction \| 0 \| \| 0001 \| AND3 \| 3 \|`	`\| 0010 0011 0100 0101 0110 \| EXACT3 2 Th3 1 EXACT3 PARITY3 NAE3 \| 2 2 2 2 2 \| \| --- \| --- \| --- \| \| 0111 \| IsomorphictoAND3 \| 3 \| \| 1000 \| IsomorphictoAND3 \| 3 \| \| 1001 1010 1011 1100 1101 \| IsomorphictoNAE3 IsomorphictoPARITY3 1 IsomorphictoEXACT3 2 IsomorphictoTh3 2 IsomorphictoEXACT \| 2 2 2 2 2 \| \| 1110 \| IsomorphictoAND3 \| 3 \| \| 1111 \| Constantfunction \| 0 \|`	`1`

Loss: BinaryCrossEntropyLoss with these parameters:

{
    "activation_fn": "torch.nn.modules.linear.Identity",
    "pos_weight": null
}

Evaluation Dataset

table-continuation-dataset

Dataset: table-continuation-dataset at 45a5089
Size: 1,466 evaluation samples
Columns: premise, hypothesis, and label

Approximate statistics based on the first 1000 samples:

	premise	hypothesis	label
type	string	string	int
details	min: 26 characters mean: 263.77 characters max: 4425 characters	min: 17 characters mean: 350.83 characters max: 3969 characters	0: ~47.70% 1: ~52.30%

Samples:

premise	hypothesis	label
`\| K40Specs \| \| \| --- \| --- \| \| GlobalMemory \| 11520MB \| \| L2cachesize \| 1.57MB \| \| Sharedmemoryperblock \| 0.049MB \|`	`\| Maxsharedmemoryperthreadblock \| 49152bytes \| \| --- \| --- \| \| L2cachesize \| 1.50MiB \|`	`0`
`\| method \| mean() \| median() \| trimean() \| \| --- \| --- \| --- \| --- \| \| Originallyreportedresults \| \| \| \| \| Shades-of-Gray \| 11.55 \| 9.70 \| 10.23 \| \| GeneralGray-World \| 11.55 \| 9.70 \| 10.23 \| \| 1st-orderGray-Edge \| 10.58 \| 8.84 \| 9.18 \| \| 2nd-orderGray-Edge \| 10.68 \| 9.02 \| 9.40 \| \| Revisitedresults \| \| \| \| \| Shades-of-Gray \| 13.32 \| 11.57 \| 12.10 \| \| GeneralGray-World \| 13.69 \| 12.11 \| 12.55 \| \| 1st-orderGray-Edge \| 11.06 \| 9.54 \| 9.81 \| \| 2nd-orderGray-Edge \| 10.73 \| 9.21 \| 9.49 \|`	`\| Greenstabilityassumptionresults \| \| \| \| \| --- \| --- \| --- \| --- \| \| Shades-of-Gray \| 12.68 \| 10.50 \| 11.25 \| \| GeneralGray-World \| 12.68 \| 10.50 \| 11.25 \| \| 1st-orderGray-Edge \| 13.41 \| 11.04 \| 11.87 \| \| 2nd-orderGray-Edge \| 12.83 \| 10.70 \| 11.44 \|`	`1`
`\| Name \| Abbreviation \| #ofinstances \| \| --- \| --- \| --- \| \| wcsp/spot5/dir \| wcsp-dir \| 21 \| \| wcsp/spot5/log \| wcsp-log \| 21 \| \| haplotyping-pedigrees \| HT \| 100 \| \| upgradeability-problem \| UP \| 100 \| \| preferenceplanning \| PP \| 29 \|`	`\| packup-wpms \| PWPMS \| 99 \| \| --- \| --- \| --- \| \| timetabling \| TT \| 26 \|`	`1`

Loss: BinaryCrossEntropyLoss with these parameters:

{
    "activation_fn": "torch.nn.modules.linear.Identity",
    "pos_weight": null
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
learning_rate: 2e-05
weight_decay: 0.01
lr_scheduler_type: cosine
warmup_ratio: 0.1
fp16: True
load_best_model_at_end: True

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
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 2e-05
weight_decay: 0.01
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 3
max_steps: -1
lr_scheduler_type: cosine
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
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
bf16: False
fp16: True
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
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: True
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
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
adafactor: False
group_by_length: False
length_column_name: length
project: huggingface
trackio_space_id: trackio
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_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
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: batch_sampler
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss	Validation Loss	eval_average_precision
-1	-1	-	-	0.5215
0.0012	1	1.568	-	-
0.0121	10	1.4514	-	-
0.0242	20	1.159	-	-
0.0364	30	1.0636	-	-
0.0485	40	1.0502	-	-
0.0606	50	0.9466	-	-
0.0727	60	0.8498	-	-
0.0848	70	0.79	-	-
0.0970	80	0.6811	-	-
0.1091	90	0.7034	-	-
0.1212	100	0.7514	-	-
0.1333	110	0.6787	-	-
0.1455	120	0.7724	-	-
0.1576	130	0.7984	-	-
0.1697	140	0.7357	-	-
0.1818	150	0.682	-	-
0.1939	160	0.7078	-	-
0.2061	170	0.735	-	-
0.2182	180	0.6515	-	-
0.2303	190	0.6872	-	-
0.2424	200	0.6303	0.6266	0.7346
0.2545	210	0.6187	-	-
0.2667	220	0.6414	-	-
0.2788	230	0.6559	-	-
0.2909	240	0.6741	-	-
0.3030	250	0.5765	-	-
0.3152	260	0.5634	-	-
0.3273	270	0.5529	-	-
0.3394	280	0.5485	-	-
0.3515	290	0.5848	-	-
0.3636	300	0.6838	-	-
0.3758	310	0.5737	-	-
0.3879	320	0.6061	-	-
0.4	330	0.559	-	-
0.4121	340	0.5933	-	-
0.4242	350	0.5114	-	-
0.4364	360	0.5575	-	-
0.4485	370	0.5679	-	-
0.4606	380	0.5434	-	-
0.4727	390	0.5161	-	-
0.4848	400	0.5138	0.5051	0.8428
0.4970	410	0.3921	-	-
0.5091	420	0.4819	-	-
0.5212	430	0.4976	-	-
0.5333	440	0.4948	-	-
0.5455	450	0.4624	-	-
0.5576	460	0.4229	-	-
0.5697	470	0.667	-	-
0.5818	480	0.541	-	-
0.5939	490	0.4705	-	-
0.6061	500	0.4292	-	-
0.6182	510	0.3875	-	-
0.6303	520	0.3439	-	-
0.6424	530	0.4438	-	-
0.6545	540	0.3549	-	-
0.6667	550	0.3269	-	-
0.6788	560	0.4639	-	-
0.6909	570	0.4241	-	-
0.7030	580	0.4069	-	-
0.7152	590	0.4362	-	-
0.7273	600	0.3688	0.4073	0.8647
0.7394	610	0.4858	-	-
0.7515	620	0.2726	-	-
0.7636	630	0.5526	-	-
0.7758	640	0.3979	-	-
0.7879	650	0.5342	-	-
0.8	660	0.3759	-	-
0.8121	670	0.3175	-	-
0.8242	680	0.454	-	-
0.8364	690	0.299	-	-
0.8485	700	0.4663	-	-
0.8606	710	0.4454	-	-
0.8727	720	0.3191	-	-
0.8848	730	0.3407	-	-
0.8970	740	0.4034	-	-
0.9091	750	0.3208	-	-
0.9212	760	0.4327	-	-
0.9333	770	0.4507	-	-
0.9455	780	0.364	-	-
0.9576	790	0.3415	-	-
0.9697	800	0.3369	0.3427	0.8937
0.9818	810	0.373	-	-
0.9939	820	0.4163	-	-
1.0061	830	0.2913	-	-
1.0182	840	0.2541	-	-
1.0303	850	0.4237	-	-
1.0424	860	0.3699	-	-
1.0545	870	0.3385	-	-
1.0667	880	0.287	-	-
1.0788	890	0.3147	-	-
1.0909	900	0.4225	-	-
1.1030	910	0.3251	-	-
1.1152	920	0.2791	-	-
1.1273	930	0.2806	-	-
1.1394	940	0.3276	-	-
1.1515	950	0.3114	-	-
1.1636	960	0.2844	-	-
1.1758	970	0.2789	-	-
1.1879	980	0.3924	-	-
1.2	990	0.2539	-	-
1.2121	1000	0.3521	0.3340	0.8968
1.2242	1010	0.3257	-	-
1.2364	1020	0.2742	-	-
1.2485	1030	0.272	-	-
1.2606	1040	0.3353	-	-
1.2727	1050	0.308	-	-
1.2848	1060	0.3421	-	-
1.2970	1070	0.3217	-	-
1.3091	1080	0.3378	-	-
1.3212	1090	0.3475	-	-
1.3333	1100	0.2731	-	-
1.3455	1110	0.2564	-	-
1.3576	1120	0.3	-	-
1.3697	1130	0.3451	-	-
1.3818	1140	0.307	-	-
1.3939	1150	0.2309	-	-
1.4061	1160	0.2663	-	-
1.4182	1170	0.267	-	-
1.4303	1180	0.2899	-	-
1.4424	1190	0.369	-	-
1.4545	1200	0.2506	0.3054	0.9073
1.4667	1210	0.3266	-	-
1.4788	1220	0.3361	-	-
1.4909	1230	0.2657	-	-
1.5030	1240	0.3517	-	-
1.5152	1250	0.289	-	-
1.5273	1260	0.2668	-	-
1.5394	1270	0.3482	-	-
1.5515	1280	0.3758	-	-
1.5636	1290	0.232	-	-
1.5758	1300	0.3564	-	-
1.5879	1310	0.2815	-	-
1.6	1320	0.2122	-	-
1.6121	1330	0.2786	-	-
1.6242	1340	0.3978	-	-
1.6364	1350	0.2222	-	-
1.6485	1360	0.2901	-	-
1.6606	1370	0.4247	-	-
1.6727	1380	0.4094	-	-
1.6848	1390	0.3077	-	-
1.6970	1400	0.2155	0.3060	0.9122
1.7091	1410	0.2607	-	-
1.7212	1420	0.2837	-	-
1.7333	1430	0.3198	-	-
1.7455	1440	0.2362	-	-
1.7576	1450	0.2265	-	-
1.7697	1460	0.3386	-	-
1.7818	1470	0.3089	-	-
1.7939	1480	0.2792	-	-
1.8061	1490	0.3103	-	-
1.8182	1500	0.364	-	-
1.8303	1510	0.2771	-	-
1.8424	1520	0.3449	-	-
1.8545	1530	0.2851	-	-
1.8667	1540	0.2513	-	-
1.8788	1550	0.3013	-	-
1.8909	1560	0.3173	-	-
1.9030	1570	0.3125	-	-
1.9152	1580	0.2399	-	-
1.9273	1590	0.2614	-	-
1.9394	1600	0.1946	0.3172	0.9165
1.9515	1610	0.2345	-	-
1.9636	1620	0.2219	-	-
1.9758	1630	0.3347	-	-
1.9879	1640	0.2964	-	-
2.0	1650	0.3139	-	-
2.0121	1660	0.2288	-	-
2.0242	1670	0.2273	-	-
2.0364	1680	0.1789	-	-
2.0485	1690	0.2484	-	-
2.0606	1700	0.1882	-	-
2.0727	1710	0.2455	-	-
2.0848	1720	0.1766	-	-
2.0970	1730	0.3073	-	-
2.1091	1740	0.2296	-	-
2.1212	1750	0.1857	-	-
2.1333	1760	0.1965	-	-
2.1455	1770	0.2159	-	-
2.1576	1780	0.1821	-	-
2.1697	1790	0.2312	-	-
2.1818	1800	0.2987	0.2863	0.9292
2.1939	1810	0.2137	-	-
2.2061	1820	0.132	-	-
2.2182	1830	0.17	-	-
2.2303	1840	0.2847	-	-
2.2424	1850	0.2753	-	-
2.2545	1860	0.2074	-	-
2.2667	1870	0.2293	-	-
2.2788	1880	0.2452	-	-
2.2909	1890	0.2294	-	-
2.3030	1900	0.238	-	-
2.3152	1910	0.1935	-	-
2.3273	1920	0.3317	-	-
2.3394	1930	0.1987	-	-
2.3515	1940	0.2472	-	-
2.3636	1950	0.22	-	-
2.3758	1960	0.1734	-	-
2.3879	1970	0.2388	-	-
2.4	1980	0.1625	-	-
2.4121	1990	0.1996	-	-
2.4242	2000	0.2388	0.2907	0.9266
2.4364	2010	0.1535	-	-
2.4485	2020	0.1842	-	-
2.4606	2030	0.3067	-	-
2.4727	2040	0.1868	-	-
2.4848	2050	0.1702	-	-
2.4970	2060	0.1749	-	-
2.5091	2070	0.1973	-	-
2.5212	2080	0.2935	-	-
2.5333	2090	0.1217	-	-
2.5455	2100	0.2526	-	-
2.5576	2110	0.223	-	-
2.5697	2120	0.1459	-	-
2.5818	2130	0.1422	-	-
2.5939	2140	0.2888	-	-
2.6061	2150	0.185	-	-
2.6182	2160	0.3221	-	-
2.6303	2170	0.2957	-	-
2.6424	2180	0.2754	-	-
2.6545	2190	0.2215	-	-
2.6667	2200	0.1802	0.2913	0.9280
2.6788	2210	0.2696	-	-
2.6909	2220	0.3014	-	-
2.7030	2230	0.1532	-	-
2.7152	2240	0.2233	-	-
2.7273	2250	0.233	-	-
2.7394	2260	0.166	-	-
2.7515	2270	0.1994	-	-
2.7636	2280	0.1746	-	-
2.7758	2290	0.1785	-	-
2.7879	2300	0.28	-	-
2.8	2310	0.2552	-	-
2.8121	2320	0.2295	-	-
2.8242	2330	0.2289	-	-
2.8364	2340	0.1959	-	-
2.8485	2350	0.223	-	-
2.8606	2360	0.1539	-	-
2.8727	2370	0.1547	-	-
2.8848	2380	0.2855	-	-
2.8970	2390	0.1785	-	-
2.9091	2400	0.1995	0.2882	0.9278
2.9212	2410	0.2055	-	-
2.9333	2420	0.1257	-	-
2.9455	2430	0.1758	-	-
2.9576	2440	0.1946	-	-
2.9697	2450	0.2797	-	-
2.9818	2460	0.2324	-	-
2.9939	2470	0.1811	-	-
-1	-1	-	-	0.9266

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.12.12
Sentence Transformers: 5.1.2
Transformers: 4.57.3
PyTorch: 2.9.0+cu126
Accelerate: 1.12.0
Datasets: 4.0.0
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",
}

Downloads last month: 10

Safetensors

Model size

32M params

Tensor type

F32

Inference Providers NEW

Text Ranking

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for BioMike/table-linker-32m-init

Base model

jhu-clsp/ettin-encoder-32m

Finetuned

(13)

this model

Dataset used to train BioMike/table-linker-32m-init

Evaluation results

Accuracy on eval
self-reported

0.894
Accuracy Threshold on eval
self-reported

0.450
F1 on eval
self-reported

0.906
F1 Threshold on eval
self-reported

0.116
Precision on eval
self-reported

0.839
Recall on eval
self-reported

0.986
Average Precision on eval
self-reported

0.927