rahulseetharaman/reranker-msmarco-v1.1-ettin-encoder-150m-listnet

CrossEncoder based on jhu-clsp/ettin-encoder-150m

This is a Cross Encoder model finetuned from jhu-clsp/ettin-encoder-150m on the ms_marco 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-150m
• Maximum Sequence Length: 7999 tokens
• Number of Output Labels: 1 label
• Training Dataset:
  • ms_marco
• Language: en

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("rahulseetharaman/reranker-msmarco-v1.1-ettin-encoder-150m-listnet")
# Get scores for pairs of texts
pairs = [
    ['how to transfer credit etisalat', 'Transferring credit from your mobile to another mobile is easy and quick. You can now transfer credit in 2 ways—Normal Credit Transfer and the Secure Credit Transfer. To transfer credit with Normal Credit Transfer, simply type: *100*Mobile Number*Amount to be transferred# then press SEND. E.g.: If you wish to send AED 5, simply type *100*05XXXXXXXX*5# then press SEND. Secure Credit Transfer Method. 5% of the transferred amount will be deducted from the sender as a transaction fee. Example: If Mobile A transferred AED 100 to Mobile B, then AED 105 (AED 100 + AED 5 = [100*5%]) will be deducted from Mobile A and Mobile B will receive AED 100.'],
    ['how to transfer credit etisalat', 'How to transfer credit on Etisalat Nigeria SIM Card. You can electronically transfer airtime from the existing credit in your Etisalat prepaid account to another Etisalat prepaid account. You can transfer any amount. MSISDN = The Etisalat number you want to transfer airtime to. Etisalat Nigeria notifies the sender and receiver about the transaction via USSD and SMS message. In case you forget your password or the SIM is locked due to entry of wrong password several times, just call Etisalat customer care.'],
    ['how to transfer credit etisalat', "Type in the number to which you want to transfer credit. Make sure the number is in the international format -- it should start with a +, followed by the two-digit country code and then the cell phone number without the first zero. Send the SMS to 1700 and wait for a message telling you how to proceed. 1 This is deducted from the sender's account. 2  You can only transfer credit from and into an active account. 3  You can only make three international credit transfers a week and a maximum of 10 per month. 4  Credit limits are set at 300 AED per transfer to a maximum of 500 AED a month."],
    ['how to transfer credit etisalat', "If you have an Etisalat prepaid or postpaid cell phone, you can transfer credit from your account to other network users. You can send credit in United Arab Emirates dirhams or in other currencies if you use the international transfer service. 1 This is deducted from the sender's account. 2  You can only transfer credit from and into an active account. 3  You can only make three international credit transfers a week and a maximum of 10 per month. 4  Credit limits are set at 300 AED per transfer to a maximum of 500 AED a month."],
    ['how to transfer credit etisalat', "Published. Etisalat has launched the ‘International Credit Transfer' facilitating its ‘Wasel’ and ‘Ahlan’ prepaid customers to transfer airtime to any prepaid account globally. Etisalat customers currently are able to transfer balance between local mobile accounts. "],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'how to transfer credit etisalat',
    [
        'Transferring credit from your mobile to another mobile is easy and quick. You can now transfer credit in 2 ways—Normal Credit Transfer and the Secure Credit Transfer. To transfer credit with Normal Credit Transfer, simply type: *100*Mobile Number*Amount to be transferred# then press SEND. E.g.: If you wish to send AED 5, simply type *100*05XXXXXXXX*5# then press SEND. Secure Credit Transfer Method. 5% of the transferred amount will be deducted from the sender as a transaction fee. Example: If Mobile A transferred AED 100 to Mobile B, then AED 105 (AED 100 + AED 5 = [100*5%]) will be deducted from Mobile A and Mobile B will receive AED 100.',
        'How to transfer credit on Etisalat Nigeria SIM Card. You can electronically transfer airtime from the existing credit in your Etisalat prepaid account to another Etisalat prepaid account. You can transfer any amount. MSISDN = The Etisalat number you want to transfer airtime to. Etisalat Nigeria notifies the sender and receiver about the transaction via USSD and SMS message. In case you forget your password or the SIM is locked due to entry of wrong password several times, just call Etisalat customer care.',
        "Type in the number to which you want to transfer credit. Make sure the number is in the international format -- it should start with a +, followed by the two-digit country code and then the cell phone number without the first zero. Send the SMS to 1700 and wait for a message telling you how to proceed. 1 This is deducted from the sender's account. 2  You can only transfer credit from and into an active account. 3  You can only make three international credit transfers a week and a maximum of 10 per month. 4  Credit limits are set at 300 AED per transfer to a maximum of 500 AED a month.",
        "If you have an Etisalat prepaid or postpaid cell phone, you can transfer credit from your account to other network users. You can send credit in United Arab Emirates dirhams or in other currencies if you use the international transfer service. 1 This is deducted from the sender's account. 2  You can only transfer credit from and into an active account. 3  You can only make three international credit transfers a week and a maximum of 10 per month. 4  Credit limits are set at 300 AED per transfer to a maximum of 500 AED a month.",
        "Published. Etisalat has launched the ‘International Credit Transfer' facilitating its ‘Wasel’ and ‘Ahlan’ prepaid customers to transfer airtime to any prepaid account globally. Etisalat customers currently are able to transfer balance between local mobile accounts. ",
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Evaluation

Metrics

Cross Encoder Reranking

• Datasets: NanoMSMARCO_R100, NanoNFCorpus_R100 and NanoNQ_R100
• Evaluated with CrossEncoderRerankingEvaluator with these parameters:

  {
      "at_k": 10,
      "always_rerank_positives": true
  }
  

Metric	NanoMSMARCO_R100	NanoNFCorpus_R100	NanoNQ_R100
map	0.5325 (+0.0429)	0.3438 (+0.0828)	0.5945 (+0.1749)
mrr@10	0.5238 (+0.0463)	0.5468 (+0.0470)	0.6074 (+0.1807)
ndcg@10	0.5898 (+0.0494)	0.3760 (+0.0510)	0.6450 (+0.1443)

Cross Encoder Nano BEIR

• Dataset: NanoBEIR_R100_mean
• Evaluated with CrossEncoderNanoBEIREvaluator with these parameters:

  {
      "dataset_names": [
          "msmarco",
          "nfcorpus",
          "nq"
      ],
      "rerank_k": 100,
      "at_k": 10,
      "always_rerank_positives": true
  }
  

Metric	Value
map	0.4903 (+0.1002)
mrr@10	0.5593 (+0.0913)
ndcg@10	0.5369 (+0.0816)

Training Details

Training Dataset

ms_marco

• Dataset: ms_marco at a47ee7a
• Size: 78,704 training samples
• Columns: query, docs, and labels
• Approximate statistics based on the first 1000 samples:

  	query	docs	labels
  type	string	list	list
  details	min: 10 characters mean: 34.47 characters max: 101 characters	min: 3 elements mean: 6.50 elements max: 10 elements	min: 3 elements mean: 6.50 elements max: 10 elements

• Samples:

  query	docs	labels
  is sugar made from beets	['How Beet Sugar is Made-the Basic Story. White beet sugar is made from the beets in a single process, rather than the two steps involved with cane sugar. Because the beets have come from the ground they are much dirtier than sugar cane and have to be thoroughly washed and separated from any remaining beet leaves, stones and other trash material before processing. Extraction. The processing starts by slicing the beets into thin chips.', 'The bigger discrepancy seems to be with brown sugar made from beets. Brown sugar made from beet sugar has molasses added to the refined white sugar. The molasses byproduct from beet sugar production is sold as an addition to animal feed and not used in food for human consumption. With cane sugar the brown sugar may be a less refined product as it a step in the production process. Sugar beets are grown commercially in twelve states, as a summer crop in northern states like Michigan and Minnesota and as a winter crop in warmer climates like California. S...	[1, 0, 0, 0, 0, ...]
  what gases are good insulators	['Gases with electronegative species (i.e. halogens such as chlorine) make good insulators, hence the popularity of SF6, which is not only dense (breakdown voltage is roughly proportional to density) but is mostly Fluorine, a highly electronegative element. Sulfur Hexafluoride (SF6) - Sulfur Hexafluoride is probably the most popular insulating gas, although its cost has risen dramatically recently. Hydrogen-Hydrogen gas is not a particularly good insulator (65% of air) from a breakdown voltage standpoint. Its very low viscosity and high thermal capacity make it an insulating gas of choice for high speed, high voltage machinery such as turbogenerators', 'Conductors and Insulators. In a conductor, electric current can flow freely, in an insulator it cannot. Metals such as copper typify conductors, while most non-metallic solids are said to be good insulators, having extremely high resistance to the flow of charge through them. Conductor implies that the outer electrons of the atoms are l...	[1, 0, 0, 0, 0, ...]
  concrete pouring cost	['Needless to say, installing and/or making concrete is a job better left to the pros, but if you are up for a DIY challenge, you will certainly save some dough. According to our concrete slab material estimator, the average minimum cost per square foot of a concrete slab is $1.36 and the maximum is $1.88. By comparison, the price per square foot for a cement slab is $5.50. Just know that even if you do not hire a pro, there are additional costs that come with all concrete slab projects. Supplies, such as the wood used to shape the concrete, can add $3.75 per square foot.', "Basic concrete slab cost. I'll base my price on what my actual estimate would be to install a 6 thick concrete slab for one of my customers. Base cost for a 6 thick concrete slab -- $4.00 dollars per square foot --. Take the length of the concrete slab x the width of the slab to get the. total square footage. Multiply the total square footage by 4 (dollars) to get the total price.", 'Labor. The cost to pour a concrete pad takes a big jump if you hire the pros, who charge per hour, per worker. If forming and pouring your pad takes three workers two days to complete at a cost of $30 per worker hour, it could add around $1,400 to your bill.', 'Video on the cost of concrete floors. Depending on the level of complexity, concrete floors can cost as little as $2 to $6 a square foot or be as expensive as $15 to $30 a square foot.']	[1, 0, 0, 0]

• Loss: ListNetLoss with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "mini_batch_size": 16
  }
  

Evaluation Dataset

ms_marco

• Dataset: ms_marco at a47ee7a
• Size: 1,000 evaluation samples
• Columns: query, docs, and labels
• Approximate statistics based on the first 1000 samples:

  	query	docs	labels
  type	string	list	list
  details	min: 10 characters mean: 33.89 characters max: 101 characters	min: 2 elements mean: 6.00 elements max: 10 elements	min: 2 elements mean: 6.00 elements max: 10 elements

• Samples:

  query	docs	labels
  how to transfer credit etisalat	['Transferring credit from your mobile to another mobile is easy and quick. You can now transfer credit in 2 ways—Normal Credit Transfer and the Secure Credit Transfer. To transfer credit with Normal Credit Transfer, simply type: 100Mobile NumberAmount to be transferred# then press SEND. E.g.: If you wish to send AED 5, simply type 10005XXXXXXXX5# then press SEND. Secure Credit Transfer Method. 5% of the transferred amount will be deducted from the sender as a transaction fee. Example: If Mobile A transferred AED 100 to Mobile B, then AED 105 (AED 100 + AED 5 = [100*5%]) will be deducted from Mobile A and Mobile B will receive AED 100.', 'How to transfer credit on Etisalat Nigeria SIM Card. You can electronically transfer airtime from the existing credit in your Etisalat prepaid account to another Etisalat prepaid account. You can transfer any amount. MSISDN = The Etisalat number you want to transfer airtime to. Etisalat Nigeria notifies the sender and receiver about the transact...	[1, 0, 0, 0, 0, ...]
  what is the transport system of the cell	["Simple definition of a transport system in biology: A transport system is a means by which materials are moved (' transported ') from an exchange surface or exchange surface. to cells* located throughout the organism. * Not all individual cells require all of the many different types of materials carried by a transport system.", 'In cellular biology, membrane transport refers to the collection of mechanisms that regulate the passage of solutes such as ions and small molecules through biological membranes, which are lipid bilayers that contain proteins embedded in them. ', 'Transport system in plant. Transport system in plant consists of tubes (vessels) of cells, adapted to perform the function of transport, so it is known as vascular system. There are two types of vessels in the plant 1. Xylem: transport water and mineral from roots to leaves. 2.', 'The transport system of the cell between the nucleus and the cytoplam is the endoplasmic rectiulum which is a complex system of membrane...	[1, 0, 0, 0, 0, ...]
  who coined the term hegemonic	['As a sociologic concept, the hegemonic nature of hegemonic masculinity derives from the theory of cultural hegemony, by Marxist theorist Antonio Gramsci, which analyzes the power relations among the social classes of a society. For example, Laurie argues that the hegemonic masculinity framework lends itself to a modified essentialism, wherein the achievement of masculine goals is frequently attributed to a way of thinking understood as inherent to the male psyche, and in relation to an innate disposition for homosocial bonding.', 'According to Nick Trujillo, author of “Hegemonic Masculinity on the Mound,” in which he analyzes the concept within the American sports culture, there are five particular dimensions. That the stereotypical male exercise physical force and control is first and foremost. ', 'Also, it could be used for the geopolitical and the cultural predominance of one country over others; from which was derived hegemonism, as in the idea that the Great Powers meant to esta...	[1, 0, 0, 0, 0, ...]

• Loss: ListNetLoss with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "mini_batch_size": 16
  }
  

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: 5
• seed: 12
• bf16: 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: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 5
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• 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: 12
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: True
• fp16: False
• 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}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• 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_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: False
• 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: False
• 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	NanoMSMARCO_R100_ndcg@10	NanoNFCorpus_R100_ndcg@10	NanoNQ_R100_ndcg@10	NanoBEIR_R100_mean_ndcg@10
-1	-1	-	-	0.0509 (-0.4895)	0.2468 (-0.0783)	0.0195 (-0.4811)	0.1057 (-0.3496)
0.0002	1	2.0621	-	-	-	-	-
0.0203	100	2.0831	2.0928	0.0270 (-0.5134)	0.2415 (-0.0835)	0.0767 (-0.4240)	0.1151 (-0.3403)
0.0407	200	2.0877	2.0909	0.0732 (-0.4673)	0.3028 (-0.0222)	0.1484 (-0.3522)	0.1748 (-0.2806)
0.0610	300	2.0823	2.0950	0.0456 (-0.4948)	0.2933 (-0.0318)	0.0394 (-0.4613)	0.1261 (-0.3293)
0.0813	400	2.0825	2.0948	0.0309 (-0.5095)	0.2526 (-0.0725)	0.0520 (-0.4486)	0.1118 (-0.3435)
0.1016	500	2.0877	2.0929	0.0507 (-0.4898)	0.3317 (+0.0067)	0.0590 (-0.4416)	0.1471 (-0.3082)
0.1220	600	2.0901	2.0915	0.1175 (-0.4229)	0.2624 (-0.0626)	0.0677 (-0.4329)	0.1492 (-0.3062)
0.1423	700	2.0775	2.0909	0.0970 (-0.4434)	0.2908 (-0.0343)	0.0585 (-0.4421)	0.1488 (-0.3066)
0.1626	800	2.0828	2.0904	0.1654 (-0.3751)	0.3020 (-0.0231)	0.1146 (-0.3860)	0.1940 (-0.2614)
0.1830	900	2.0773	2.0906	0.1242 (-0.4162)	0.2828 (-0.0422)	0.0940 (-0.4067)	0.1670 (-0.2884)
0.2033	1000	2.0746	2.0858	0.4331 (-0.1073)	0.3011 (-0.0239)	0.4507 (-0.0500)	0.3950 (-0.0604)
0.2236	1100	2.0764	2.0839	0.4451 (-0.0953)	0.3484 (+0.0234)	0.4370 (-0.0637)	0.4102 (-0.0452)
0.2440	1200	2.087	2.0822	0.4766 (-0.0638)	0.3698 (+0.0448)	0.4646 (-0.0361)	0.4370 (-0.0184)
0.2643	1300	2.074	2.0813	0.5038 (-0.0366)	0.3605 (+0.0355)	0.4657 (-0.0350)	0.4433 (-0.0120)
0.2846	1400	2.0768	2.0803	0.5417 (+0.0013)	0.3748 (+0.0498)	0.5447 (+0.0441)	0.4871 (+0.0317)
0.3049	1500	2.0749	2.0798	0.5016 (-0.0389)	0.3661 (+0.0411)	0.5801 (+0.0795)	0.4826 (+0.0272)
0.3253	1600	2.0803	2.0812	0.5325 (-0.0080)	0.3759 (+0.0508)	0.5822 (+0.0816)	0.4969 (+0.0415)
0.3456	1700	2.0759	2.0792	0.5462 (+0.0058)	0.3803 (+0.0552)	0.5532 (+0.0525)	0.4932 (+0.0379)
0.3659	1800	2.0795	2.0787	0.5299 (-0.0106)	0.3691 (+0.0441)	0.5743 (+0.0737)	0.4911 (+0.0357)
0.3863	1900	2.0651	2.0788	0.5356 (-0.0048)	0.3583 (+0.0332)	0.5916 (+0.0909)	0.4952 (+0.0398)
0.4066	2000	2.0719	2.0785	0.5719 (+0.0315)	0.3887 (+0.0637)	0.5866 (+0.0860)	0.5158 (+0.0604)
0.4269	2100	2.07	2.0784	0.5549 (+0.0145)	0.3516 (+0.0266)	0.6274 (+0.1267)	0.5113 (+0.0559)
0.4472	2200	2.0662	2.0778	0.5629 (+0.0224)	0.3750 (+0.0500)	0.5731 (+0.0724)	0.5036 (+0.0483)
0.4676	2300	2.0695	2.0774	0.5723 (+0.0319)	0.3479 (+0.0228)	0.6157 (+0.1150)	0.5120 (+0.0566)
0.4879	2400	2.0736	2.0777	0.5560 (+0.0155)	0.3711 (+0.0460)	0.6402 (+0.1395)	0.5224 (+0.0670)
0.5082	2500	2.0617	2.0776	0.5274 (-0.0130)	0.3585 (+0.0334)	0.6576 (+0.1570)	0.5145 (+0.0591)
0.5286	2600	2.0807	2.0770	0.5362 (-0.0043)	0.3706 (+0.0456)	0.6021 (+0.1014)	0.5029 (+0.0476)
0.5489	2700	2.0716	2.0772	0.5514 (+0.0110)	0.3849 (+0.0599)	0.6472 (+0.1466)	0.5278 (+0.0725)
0.5692	2800	2.0691	2.0767	0.5551 (+0.0146)	0.3593 (+0.0343)	0.6073 (+0.1066)	0.5072 (+0.0519)
0.5896	2900	2.0646	2.0770	0.5402 (-0.0003)	0.3882 (+0.0632)	0.5872 (+0.0865)	0.5052 (+0.0498)
0.6099	3000	2.0846	2.0777	0.5278 (-0.0126)	0.3600 (+0.0350)	0.5853 (+0.0847)	0.4911 (+0.0357)
0.6302	3100	2.0765	2.0768	0.5534 (+0.0130)	0.3734 (+0.0483)	0.6369 (+0.1363)	0.5212 (+0.0659)
0.6505	3200	2.0629	2.0767	0.5713 (+0.0309)	0.3623 (+0.0373)	0.6483 (+0.1476)	0.5273 (+0.0719)
0.6709	3300	2.0757	2.0764	0.5898 (+0.0494)	0.3760 (+0.0510)	0.6450 (+0.1443)	0.5369 (+0.0816)
0.6912	3400	2.0741	2.0765	0.5639 (+0.0234)	0.3866 (+0.0616)	0.6271 (+0.1265)	0.5259 (+0.0705)
0.7115	3500	2.0736	2.0766	0.5541 (+0.0136)	0.3833 (+0.0583)	0.6304 (+0.1298)	0.5226 (+0.0672)
0.7319	3600	2.0748	2.0763	0.5242 (-0.0163)	0.3648 (+0.0398)	0.6560 (+0.1554)	0.5150 (+0.0596)
0.7522	3700	2.0725	2.0762	0.5475 (+0.0071)	0.3561 (+0.0311)	0.5939 (+0.0933)	0.4992 (+0.0438)
0.7725	3800	2.0684	2.0761	0.5480 (+0.0076)	0.3637 (+0.0387)	0.6472 (+0.1466)	0.5196 (+0.0643)
0.7928	3900	2.0702	2.0762	0.6007 (+0.0603)	0.3442 (+0.0192)	0.6246 (+0.1240)	0.5232 (+0.0678)
0.8132	4000	2.0709	2.0767	0.5732 (+0.0327)	0.3369 (+0.0119)	0.6343 (+0.1337)	0.5148 (+0.0594)
0.8335	4100	2.0791	2.0762	0.5756 (+0.0352)	0.3397 (+0.0146)	0.6232 (+0.1226)	0.5128 (+0.0575)
0.8538	4200	2.0759	2.0760	0.5940 (+0.0536)	0.3549 (+0.0299)	0.6199 (+0.1192)	0.5230 (+0.0676)
0.8742	4300	2.0731	2.0760	0.5708 (+0.0304)	0.3610 (+0.0360)	0.6179 (+0.1173)	0.5166 (+0.0612)
0.8945	4400	2.0643	2.0757	0.5602 (+0.0198)	0.3647 (+0.0397)	0.5874 (+0.0867)	0.5041 (+0.0487)
0.9148	4500	2.0743	2.0761	0.5783 (+0.0379)	0.3556 (+0.0306)	0.6362 (+0.1356)	0.5234 (+0.0680)
0.9351	4600	2.0636	2.0758	0.5646 (+0.0241)	0.3447 (+0.0197)	0.6049 (+0.1043)	0.5047 (+0.0494)
0.9555	4700	2.0778	2.0761	0.5727 (+0.0322)	0.3384 (+0.0134)	0.6178 (+0.1171)	0.5096 (+0.0543)
0.9758	4800	2.074	2.0758	0.5670 (+0.0266)	0.3557 (+0.0307)	0.6099 (+0.1093)	0.5109 (+0.0555)
0.9961	4900	2.07	2.0759	0.5468 (+0.0063)	0.3647 (+0.0396)	0.6388 (+0.1381)	0.5167 (+0.0614)
1.0165	5000	2.0614	2.0759	0.5481 (+0.0077)	0.3526 (+0.0276)	0.6461 (+0.1455)	0.5156 (+0.0603)
1.0368	5100	2.0643	2.0759	0.5600 (+0.0196)	0.3524 (+0.0274)	0.6085 (+0.1079)	0.5070 (+0.0516)
1.0571	5200	2.0672	2.0760	0.5733 (+0.0329)	0.3545 (+0.0294)	0.6061 (+0.1055)	0.5113 (+0.0559)
1.0775	5300	2.0667	2.0759	0.5722 (+0.0318)	0.3567 (+0.0316)	0.6351 (+0.1344)	0.5213 (+0.0660)
1.0978	5400	2.0701	2.0762	0.5501 (+0.0097)	0.3713 (+0.0462)	0.6411 (+0.1405)	0.5208 (+0.0655)
1.1181	5500	2.0648	2.0765	0.5462 (+0.0058)	0.3733 (+0.0483)	0.6208 (+0.1202)	0.5134 (+0.0581)
1.1384	5600	2.0614	2.0759	0.5509 (+0.0105)	0.3903 (+0.0653)	0.5935 (+0.0928)	0.5116 (+0.0562)
1.1588	5700	2.0655	2.0764	0.5464 (+0.0060)	0.3676 (+0.0426)	0.5730 (+0.0724)	0.4957 (+0.0403)
1.1791	5800	2.0733	2.0761	0.5607 (+0.0203)	0.3465 (+0.0214)	0.6030 (+0.1024)	0.5034 (+0.0480)
1.1994	5900	2.0674	2.0765	0.5272 (-0.0133)	0.3640 (+0.0390)	0.5542 (+0.0536)	0.4818 (+0.0264)
1.2198	6000	2.0737	2.0761	0.5599 (+0.0194)	0.3498 (+0.0247)	0.5943 (+0.0936)	0.5013 (+0.0459)
1.2401	6100	2.0618	2.0763	0.5579 (+0.0174)	0.3633 (+0.0382)	0.5856 (+0.0850)	0.5022 (+0.0469)
1.2604	6200	2.0621	2.0762	0.5879 (+0.0475)	0.3771 (+0.0521)	0.6114 (+0.1108)	0.5255 (+0.0701)
1.2807	6300	2.0803	2.0760	0.5521 (+0.0117)	0.3513 (+0.0263)	0.6230 (+0.1223)	0.5088 (+0.0534)
1.3011	6400	2.059	2.0759	0.5739 (+0.0335)	0.3582 (+0.0331)	0.6165 (+0.1159)	0.5162 (+0.0608)
1.3214	6500	2.0708	2.0759	0.5218 (-0.0186)	0.3348 (+0.0098)	0.6288 (+0.1281)	0.4951 (+0.0398)
1.3417	6600	2.0584	2.0761	0.5522 (+0.0118)	0.3656 (+0.0405)	0.6456 (+0.1449)	0.5211 (+0.0657)
1.3621	6700	2.0581	2.0760	0.5326 (-0.0078)	0.3504 (+0.0253)	0.6119 (+0.1112)	0.4983 (+0.0429)
1.3824	6800	2.0618	2.0758	0.5589 (+0.0184)	0.3695 (+0.0445)	0.6101 (+0.1095)	0.5128 (+0.0575)
1.4027	6900	2.0576	2.0765	0.5571 (+0.0167)	0.3624 (+0.0373)	0.6230 (+0.1224)	0.5142 (+0.0588)
1.4231	7000	2.0652	2.0759	0.5354 (-0.0050)	0.3579 (+0.0328)	0.6062 (+0.1056)	0.4998 (+0.0445)
1.4434	7100	2.0646	2.0758	0.5316 (-0.0088)	0.3572 (+0.0321)	0.6205 (+0.1198)	0.5031 (+0.0477)
1.4637	7200	2.0701	2.0758	0.5367 (-0.0037)	0.3679 (+0.0428)	0.6702 (+0.1696)	0.5249 (+0.0696)
1.4840	7300	2.0751	2.0759	0.5430 (+0.0026)	0.3586 (+0.0335)	0.5973 (+0.0967)	0.4996 (+0.0443)
1.5044	7400	2.0661	2.0764	0.5827 (+0.0423)	0.3881 (+0.0631)	0.6341 (+0.1334)	0.5350 (+0.0796)
1.5247	7500	2.0742	2.0757	0.5416 (+0.0012)	0.3907 (+0.0656)	0.5900 (+0.0894)	0.5074 (+0.0521)
1.5450	7600	2.0599	2.0765	0.5336 (-0.0068)	0.3682 (+0.0432)	0.5989 (+0.0983)	0.5002 (+0.0449)
1.5654	7700	2.0719	2.0759	0.5402 (-0.0002)	0.3661 (+0.0411)	0.6070 (+0.1063)	0.5044 (+0.0491)
1.5857	7800	2.0715	2.0756	0.5590 (+0.0186)	0.3612 (+0.0362)	0.6180 (+0.1173)	0.5127 (+0.0574)
1.6060	7900	2.0686	2.0759	0.5665 (+0.0260)	0.3518 (+0.0267)	0.5853 (+0.0846)	0.5012 (+0.0458)
1.6263	8000	2.0597	2.0757	0.5646 (+0.0242)	0.3662 (+0.0412)	0.6050 (+0.1043)	0.5119 (+0.0566)
1.6467	8100	2.072	2.0756	0.5471 (+0.0067)	0.3493 (+0.0243)	0.6150 (+0.1143)	0.5038 (+0.0484)
1.6670	8200	2.0621	2.0755	0.5536 (+0.0132)	0.3415 (+0.0164)	0.6082 (+0.1076)	0.5011 (+0.0457)
1.6873	8300	2.0653	2.0757	0.5585 (+0.0180)	0.3267 (+0.0016)	0.6198 (+0.1192)	0.5016 (+0.0463)
1.7077	8400	2.0679	2.0754	0.5813 (+0.0409)	0.3346 (+0.0095)	0.6197 (+0.1190)	0.5119 (+0.0565)
1.7280	8500	2.0663	2.0759	0.5619 (+0.0215)	0.3482 (+0.0232)	0.6041 (+0.1035)	0.5048 (+0.0494)
1.7483	8600	2.0717	2.0759	0.5663 (+0.0258)	0.3354 (+0.0104)	0.6087 (+0.1080)	0.5035 (+0.0481)
1.7687	8700	2.0584	2.0758	0.5450 (+0.0046)	0.3439 (+0.0188)	0.6328 (+0.1321)	0.5072 (+0.0519)
1.7890	8800	2.0645	2.0756	0.5244 (-0.0160)	0.3479 (+0.0229)	0.6517 (+0.1510)	0.5080 (+0.0526)
1.8093	8900	2.0621	2.0762	0.5394 (-0.0010)	0.3302 (+0.0051)	0.6306 (+0.1299)	0.5001 (+0.0447)
1.8296	9000	2.0634	2.0757	0.5792 (+0.0388)	0.3254 (+0.0004)	0.6016 (+0.1009)	0.5021 (+0.0467)
1.8500	9100	2.0697	2.0758	0.5571 (+0.0167)	0.3307 (+0.0057)	0.6091 (+0.1085)	0.4990 (+0.0436)
1.8703	9200	2.066	2.0756	0.5550 (+0.0145)	0.3467 (+0.0217)	0.6023 (+0.1017)	0.5013 (+0.0460)
1.8906	9300	2.0612	2.0760	0.5541 (+0.0137)	0.3464 (+0.0214)	0.6126 (+0.1120)	0.5044 (+0.0490)
1.9110	9400	2.0678	2.0752	0.5349 (-0.0055)	0.3585 (+0.0335)	0.6248 (+0.1241)	0.5061 (+0.0507)
1.9313	9500	2.0715	2.0756	0.5459 (+0.0054)	0.3613 (+0.0362)	0.6308 (+0.1301)	0.5126 (+0.0573)
1.9516	9600	2.0586	2.0753	0.5346 (-0.0059)	0.3583 (+0.0332)	0.6446 (+0.1439)	0.5125 (+0.0571)
1.9719	9700	2.0573	2.0759	0.5462 (+0.0058)	0.3551 (+0.0301)	0.6515 (+0.1509)	0.5176 (+0.0623)
1.9923	9800	2.0634	2.0753	0.5465 (+0.0061)	0.3644 (+0.0394)	0.6047 (+0.1040)	0.5052 (+0.0498)
2.0126	9900	2.0659	2.0766	0.5366 (-0.0038)	0.3529 (+0.0279)	0.5920 (+0.0914)	0.4938 (+0.0385)
2.0329	10000	2.0514	2.0780	0.5466 (+0.0062)	0.3489 (+0.0239)	0.5667 (+0.0660)	0.4874 (+0.0320)
2.0533	10100	2.0506	2.0782	0.5444 (+0.0040)	0.3564 (+0.0313)	0.5532 (+0.0525)	0.4847 (+0.0293)
2.0736	10200	2.0528	2.0777	0.5340 (-0.0065)	0.3289 (+0.0038)	0.5812 (+0.0805)	0.4813 (+0.0260)
2.0939	10300	2.0552	2.0781	0.5045 (-0.0359)	0.3402 (+0.0152)	0.5500 (+0.0494)	0.4649 (+0.0096)
2.1143	10400	2.0551	2.0798	0.5110 (-0.0294)	0.3490 (+0.0240)	0.5577 (+0.0570)	0.4726 (+0.0172)
2.1346	10500	2.0472	2.0782	0.5238 (-0.0166)	0.3567 (+0.0316)	0.5681 (+0.0674)	0.4829 (+0.0275)
2.1549	10600	2.0568	2.0796	0.5299 (-0.0106)	0.3561 (+0.0311)	0.5770 (+0.0763)	0.4876 (+0.0323)
2.1752	10700	2.059	2.0794	0.5300 (-0.0105)	0.3435 (+0.0184)	0.5440 (+0.0434)	0.4725 (+0.0171)
2.1956	10800	2.0608	2.0783	0.5444 (+0.0040)	0.3497 (+0.0246)	0.5827 (+0.0820)	0.4923 (+0.0369)
2.2159	10900	2.0544	2.0792	0.5239 (-0.0165)	0.3334 (+0.0083)	0.5381 (+0.0374)	0.4651 (+0.0097)
2.2362	11000	2.0581	2.0784	0.5227 (-0.0177)	0.3410 (+0.0159)	0.5669 (+0.0662)	0.4769 (+0.0215)
2.2566	11100	2.0503	2.0798	0.4970 (-0.0435)	0.3363 (+0.0112)	0.5674 (+0.0667)	0.4669 (+0.0115)
2.2769	11200	2.0587	2.0783	0.5088 (-0.0317)	0.3405 (+0.0155)	0.5607 (+0.0601)	0.4700 (+0.0146)
2.2972	11300	2.058	2.0805	0.4763 (-0.0641)	0.3391 (+0.0140)	0.5457 (+0.0451)	0.4537 (-0.0017)
2.3175	11400	2.0481	2.0803	0.4811 (-0.0593)	0.3442 (+0.0192)	0.5410 (+0.0403)	0.4554 (+0.0001)
2.3379	11500	2.0507	2.0791	0.5030 (-0.0374)	0.3337 (+0.0086)	0.5680 (+0.0673)	0.4682 (+0.0129)
2.3582	11600	2.055	2.0792	0.4964 (-0.0440)	0.3405 (+0.0154)	0.5717 (+0.0711)	0.4695 (+0.0142)
2.3785	11700	2.0635	2.0815	0.4840 (-0.0564)	0.3492 (+0.0241)	0.5836 (+0.0830)	0.4723 (+0.0169)
2.3989	11800	2.054	2.0793	0.4744 (-0.0660)	0.3428 (+0.0177)	0.5253 (+0.0247)	0.4475 (-0.0079)
2.4192	11900	2.0489	2.0786	0.4984 (-0.0420)	0.3333 (+0.0083)	0.5213 (+0.0207)	0.4510 (-0.0044)
2.4395	12000	2.0592	2.0794	0.4884 (-0.0520)	0.3586 (+0.0336)	0.5415 (+0.0409)	0.4629 (+0.0075)
2.4598	12100	2.0414	2.0808	0.4966 (-0.0438)	0.3583 (+0.0332)	0.5937 (+0.0931)	0.4829 (+0.0275)
2.4802	12200	2.0594	2.0808	0.5062 (-0.0342)	0.3423 (+0.0173)	0.5675 (+0.0669)	0.4720 (+0.0167)
2.5005	12300	2.0473	2.0791	0.5141 (-0.0263)	0.3357 (+0.0106)	0.5754 (+0.0748)	0.4751 (+0.0197)
2.5208	12400	2.0587	2.0797	0.5011 (-0.0393)	0.3528 (+0.0278)	0.6013 (+0.1007)	0.4851 (+0.0297)
2.5412	12500	2.0647	2.0789	0.5021 (-0.0384)	0.3519 (+0.0269)	0.5815 (+0.0809)	0.4785 (+0.0231)
2.5615	12600	2.0436	2.0798	0.5156 (-0.0248)	0.3509 (+0.0259)	0.5845 (+0.0839)	0.4837 (+0.0283)
2.5818	12700	2.0543	2.0796	0.5150 (-0.0255)	0.3541 (+0.0291)	0.5852 (+0.0845)	0.4848 (+0.0294)
2.6022	12800	2.0562	2.0799	0.5079 (-0.0325)	0.3605 (+0.0354)	0.5974 (+0.0967)	0.4886 (+0.0332)
2.6225	12900	2.0609	2.0802	0.5194 (-0.0210)	0.3544 (+0.0293)	0.5770 (+0.0763)	0.4836 (+0.0282)
2.6428	13000	2.0493	2.0810	0.5179 (-0.0225)	0.3536 (+0.0285)	0.5621 (+0.0614)	0.4778 (+0.0225)
2.6631	13100	2.0497	2.0806	0.5102 (-0.0302)	0.3561 (+0.0311)	0.5521 (+0.0515)	0.4728 (+0.0175)
2.6835	13200	2.048	2.0789	0.5208 (-0.0196)	0.3557 (+0.0307)	0.5507 (+0.0501)	0.4758 (+0.0204)
2.7038	13300	2.0496	2.0788	0.4947 (-0.0458)	0.3525 (+0.0274)	0.5395 (+0.0388)	0.4622 (+0.0068)
2.7241	13400	2.0595	2.0784	0.4846 (-0.0558)	0.3533 (+0.0283)	0.5264 (+0.0257)	0.4548 (-0.0006)
2.7445	13500	2.0469	2.0797	0.4873 (-0.0531)	0.3559 (+0.0308)	0.5500 (+0.0494)	0.4644 (+0.0090)
2.7648	13600	2.0502	2.0787	0.4928 (-0.0476)	0.3534 (+0.0284)	0.5608 (+0.0601)	0.4690 (+0.0137)
2.7851	13700	2.062	2.0790	0.4826 (-0.0578)	0.3629 (+0.0378)	0.5539 (+0.0532)	0.4665 (+0.0111)
2.8054	13800	2.0583	2.0796	0.4838 (-0.0566)	0.3714 (+0.0464)	0.5506 (+0.0499)	0.4686 (+0.0132)
2.8258	13900	2.0563	2.0802	0.5056 (-0.0348)	0.3573 (+0.0323)	0.5560 (+0.0554)	0.4730 (+0.0176)
2.8461	14000	2.0624	2.0792	0.5001 (-0.0404)	0.3549 (+0.0299)	0.5459 (+0.0453)	0.4670 (+0.0116)
2.8664	14100	2.0623	2.0790	0.5096 (-0.0309)	0.3741 (+0.0491)	0.5426 (+0.0419)	0.4754 (+0.0200)
2.8868	14200	2.0609	2.0792	0.5122 (-0.0282)	0.3730 (+0.0479)	0.5261 (+0.0255)	0.4704 (+0.0151)
2.9071	14300	2.0522	2.0789	0.5244 (-0.0160)	0.3727 (+0.0476)	0.5549 (+0.0542)	0.4840 (+0.0286)
2.9274	14400	2.0576	2.0790	0.5215 (-0.0189)	0.3668 (+0.0417)	0.5466 (+0.0460)	0.4783 (+0.0229)
2.9478	14500	2.0489	2.0797	0.5195 (-0.0209)	0.3699 (+0.0448)	0.5520 (+0.0513)	0.4805 (+0.0251)
2.9681	14600	2.0421	2.0788	0.5184 (-0.0221)	0.3740 (+0.0490)	0.5494 (+0.0487)	0.4806 (+0.0252)
2.9884	14700	2.0523	2.0795	0.5034 (-0.0371)	0.3734 (+0.0483)	0.5535 (+0.0528)	0.4767 (+0.0214)
3.0087	14800	2.0386	2.0816	0.5120 (-0.0284)	0.3784 (+0.0534)	0.5241 (+0.0234)	0.4715 (+0.0161)
3.0291	14900	2.0298	2.0850	0.4934 (-0.0471)	0.3687 (+0.0437)	0.5259 (+0.0253)	0.4627 (+0.0073)
3.0494	15000	2.0421	2.0843	0.4711 (-0.0694)	0.3680 (+0.0429)	0.5447 (+0.0440)	0.4612 (+0.0059)
3.0697	15100	2.0404	2.0851	0.4817 (-0.0587)	0.3647 (+0.0396)	0.5243 (+0.0237)	0.4569 (+0.0015)
3.0901	15200	2.0281	2.0855	0.4825 (-0.0579)	0.3517 (+0.0266)	0.4892 (-0.0114)	0.4411 (-0.0142)
3.1104	15300	2.0219	2.0865	0.4477 (-0.0928)	0.3636 (+0.0386)	0.5097 (+0.0090)	0.4403 (-0.0150)
3.1307	15400	2.0449	2.0858	0.4554 (-0.0850)	0.3534 (+0.0283)	0.5038 (+0.0031)	0.4375 (-0.0179)
3.1510	15500	2.0337	2.0863	0.4450 (-0.0954)	0.3547 (+0.0296)	0.5142 (+0.0135)	0.4380 (-0.0174)
3.1714	15600	2.0401	2.0867	0.4525 (-0.0879)	0.3525 (+0.0275)	0.4985 (-0.0021)	0.4345 (-0.0208)
3.1917	15700	2.0326	2.0865	0.4332 (-0.1073)	0.3420 (+0.0170)	0.5093 (+0.0087)	0.4282 (-0.0272)
3.2120	15800	2.0383	2.0853	0.4601 (-0.0803)	0.3502 (+0.0251)	0.4933 (-0.0073)	0.4346 (-0.0208)
3.2324	15900	2.0363	2.0863	0.4727 (-0.0677)	0.3458 (+0.0208)	0.4827 (-0.0180)	0.4337 (-0.0216)
3.2527	16000	2.0302	2.0845	0.4590 (-0.0814)	0.3485 (+0.0235)	0.4647 (-0.0359)	0.4241 (-0.0313)
3.2730	16100	2.0438	2.0866	0.4531 (-0.0873)	0.3596 (+0.0346)	0.4927 (-0.0080)	0.4351 (-0.0202)
3.2934	16200	2.0354	2.0848	0.4331 (-0.1073)	0.3436 (+0.0186)	0.4802 (-0.0205)	0.4190 (-0.0364)
3.3137	16300	2.036	2.0859	0.4514 (-0.0890)	0.3637 (+0.0386)	0.4919 (-0.0087)	0.4357 (-0.0197)
3.3340	16400	2.0439	2.0870	0.4486 (-0.0919)	0.3519 (+0.0269)	0.4808 (-0.0198)	0.4271 (-0.0283)
3.3543	16500	2.0346	2.0847	0.4647 (-0.0757)	0.3511 (+0.0261)	0.5037 (+0.0030)	0.4398 (-0.0155)
3.3747	16600	2.0404	2.0847	0.4494 (-0.0910)	0.3587 (+0.0337)	0.5064 (+0.0058)	0.4382 (-0.0172)
3.3950	16700	2.0283	2.0860	0.4593 (-0.0812)	0.3540 (+0.0289)	0.4988 (-0.0018)	0.4374 (-0.0180)
3.4153	16800	2.0403	2.0863	0.4439 (-0.0965)	0.3525 (+0.0274)	0.4779 (-0.0227)	0.4248 (-0.0306)
3.4357	16900	2.0396	2.0851	0.4530 (-0.0875)	0.3527 (+0.0276)	0.5003 (-0.0004)	0.4353 (-0.0201)
3.4560	17000	2.0383	2.0847	0.4314 (-0.1091)	0.3609 (+0.0358)	0.5090 (+0.0083)	0.4337 (-0.0216)
3.4763	17100	2.0291	2.0866	0.4349 (-0.1056)	0.3504 (+0.0254)	0.4976 (-0.0031)	0.4276 (-0.0277)
3.4966	17200	2.0456	2.0875	0.4407 (-0.0997)	0.3531 (+0.0280)	0.4960 (-0.0046)	0.4299 (-0.0254)
3.5170	17300	2.0372	2.0845	0.4430 (-0.0974)	0.3650 (+0.0399)	0.5088 (+0.0081)	0.4389 (-0.0164)
3.5373	17400	2.0296	2.0876	0.4554 (-0.0851)	0.3612 (+0.0361)	0.4999 (-0.0007)	0.4388 (-0.0165)
3.5576	17500	2.0449	2.0860	0.4526 (-0.0879)	0.3567 (+0.0317)	0.4899 (-0.0108)	0.4331 (-0.0223)
3.5780	17600	2.0342	2.0872	0.4590 (-0.0814)	0.3599 (+0.0348)	0.4997 (-0.0010)	0.4395 (-0.0158)
3.5983	17700	2.0458	2.0854	0.4419 (-0.0985)	0.3607 (+0.0357)	0.4933 (-0.0074)	0.4320 (-0.0234)
3.6186	17800	2.0373	2.0871	0.4441 (-0.0963)	0.3581 (+0.0331)	0.5107 (+0.0101)	0.4377 (-0.0177)
3.6390	17900	2.0377	2.0853	0.4520 (-0.0885)	0.3563 (+0.0313)	0.4944 (-0.0062)	0.4342 (-0.0211)
3.6593	18000	2.0374	2.0869	0.4268 (-0.1137)	0.3616 (+0.0365)	0.5196 (+0.0189)	0.4360 (-0.0194)
3.6796	18100	2.0365	2.0862	0.4335 (-0.1069)	0.3581 (+0.0331)	0.4978 (-0.0028)	0.4298 (-0.0256)
3.6999	18200	2.0393	2.0856	0.4448 (-0.0956)	0.3620 (+0.0370)	0.4978 (-0.0029)	0.4349 (-0.0205)
3.7203	18300	2.0277	2.0861	0.4398 (-0.1007)	0.3578 (+0.0328)	0.4823 (-0.0183)	0.4266 (-0.0287)
3.7406	18400	2.0456	2.0867	0.4402 (-0.1003)	0.3495 (+0.0245)	0.4765 (-0.0242)	0.4221 (-0.0333)
3.7609	18500	2.031	2.0871	0.4532 (-0.0872)	0.3447 (+0.0196)	0.5144 (+0.0137)	0.4374 (-0.0180)
3.7813	18600	2.0279	2.0865	0.4628 (-0.0776)	0.3474 (+0.0224)	0.5013 (+0.0007)	0.4372 (-0.0182)
3.8016	18700	2.0412	2.0849	0.4518 (-0.0887)	0.3533 (+0.0283)	0.4898 (-0.0108)	0.4316 (-0.0237)
3.8219	18800	2.0345	2.0858	0.4627 (-0.0777)	0.3467 (+0.0217)	0.4885 (-0.0121)	0.4327 (-0.0227)
3.8422	18900	2.0391	2.0858	0.4603 (-0.0801)	0.3401 (+0.0150)	0.4664 (-0.0342)	0.4223 (-0.0331)
3.8626	19000	2.0334	2.0879	0.4523 (-0.0882)	0.3504 (+0.0253)	0.5032 (+0.0025)	0.4353 (-0.0201)
3.8829	19100	2.0443	2.0870	0.4256 (-0.1148)	0.3554 (+0.0304)	0.5072 (+0.0066)	0.4294 (-0.0259)
3.9032	19200	2.0391	2.0866	0.4372 (-0.1032)	0.3538 (+0.0287)	0.4997 (-0.0009)	0.4302 (-0.0252)
3.9236	19300	2.039	2.0859	0.4392 (-0.1013)	0.3472 (+0.0221)	0.4994 (-0.0013)	0.4286 (-0.0268)
3.9439	19400	2.0344	2.0859	0.4586 (-0.0818)	0.3553 (+0.0303)	0.4800 (-0.0206)	0.4313 (-0.0240)
3.9642	19500	2.0373	2.0855	0.4460 (-0.0944)	0.3579 (+0.0329)	0.4853 (-0.0153)	0.4298 (-0.0256)
3.9845	19600	2.0334	2.0867	0.4517 (-0.0888)	0.3558 (+0.0307)	0.4866 (-0.0140)	0.4314 (-0.0240)
4.0049	19700	2.0305	2.0885	0.4414 (-0.0990)	0.3644 (+0.0394)	0.4807 (-0.0199)	0.4288 (-0.0265)
4.0252	19800	2.0246	2.0898	0.4413 (-0.0991)	0.3506 (+0.0255)	0.4561 (-0.0446)	0.4160 (-0.0394)
4.0455	19900	2.0252	2.0886	0.4442 (-0.0963)	0.3524 (+0.0274)	0.4609 (-0.0397)	0.4192 (-0.0362)
4.0659	20000	2.0244	2.0883	0.4377 (-0.1027)	0.3536 (+0.0285)	0.4624 (-0.0383)	0.4179 (-0.0375)
4.0862	20100	2.0203	2.0902	0.4245 (-0.1159)	0.3570 (+0.0320)	0.4667 (-0.0339)	0.4161 (-0.0393)
4.1065	20200	2.0254	2.0902	0.4236 (-0.1168)	0.3601 (+0.0350)	0.4730 (-0.0277)	0.4189 (-0.0365)
4.1269	20300	2.0223	2.0899	0.4231 (-0.1173)	0.3576 (+0.0326)	0.4564 (-0.0443)	0.4124 (-0.0430)
4.1472	20400	2.029	2.0906	0.4168 (-0.1237)	0.3546 (+0.0296)	0.4536 (-0.0470)	0.4083 (-0.0470)
4.1675	20500	2.0268	2.0905	0.4072 (-0.1332)	0.3577 (+0.0326)	0.4670 (-0.0336)	0.4106 (-0.0447)
4.1878	20600	2.0244	2.0913	0.4068 (-0.1336)	0.3442 (+0.0192)	0.4718 (-0.0289)	0.4076 (-0.0478)
4.2082	20700	2.0247	2.0911	0.4068 (-0.1336)	0.3459 (+0.0209)	0.4708 (-0.0298)	0.4079 (-0.0475)
4.2285	20800	2.0181	2.0900	0.4138 (-0.1266)	0.3487 (+0.0237)	0.4613 (-0.0393)	0.4079 (-0.0474)
4.2488	20900	2.0234	2.0911	0.4240 (-0.1164)	0.3530 (+0.0279)	0.4681 (-0.0326)	0.4150 (-0.0403)
4.2692	21000	2.0198	2.0901	0.4075 (-0.1329)	0.3482 (+0.0232)	0.4594 (-0.0413)	0.4050 (-0.0503)
4.2895	21100	2.0245	2.0912	0.4049 (-0.1355)	0.3390 (+0.0140)	0.4734 (-0.0273)	0.4058 (-0.0496)
4.3098	21200	2.0234	2.0900	0.4017 (-0.1387)	0.3463 (+0.0213)	0.4707 (-0.0300)	0.4062 (-0.0491)
4.3301	21300	2.0359	2.0898	0.4021 (-0.1383)	0.3485 (+0.0235)	0.4795 (-0.0211)	0.4100 (-0.0453)
4.3505	21400	2.0234	2.0898	0.4075 (-0.1329)	0.3436 (+0.0185)	0.4649 (-0.0358)	0.4053 (-0.0500)
4.3708	21500	2.0229	2.0905	0.4194 (-0.1210)	0.3442 (+0.0191)	0.4680 (-0.0326)	0.4105 (-0.0448)
4.3911	21600	2.0226	2.0896	0.4245 (-0.1159)	0.3432 (+0.0182)	0.4688 (-0.0319)	0.4122 (-0.0432)
4.4115	21700	2.0257	2.0908	0.4145 (-0.1259)	0.3414 (+0.0164)	0.4567 (-0.0440)	0.4042 (-0.0512)
4.4318	21800	2.0212	2.0902	0.4149 (-0.1255)	0.3487 (+0.0236)	0.4550 (-0.0457)	0.4062 (-0.0492)
4.4521	21900	2.0281	2.0916	0.3951 (-0.1453)	0.3453 (+0.0203)	0.4601 (-0.0405)	0.4002 (-0.0552)
4.4725	22000	2.0312	2.0909	0.4023 (-0.1382)	0.3459 (+0.0208)	0.4570 (-0.0437)	0.4017 (-0.0537)
4.4928	22100	2.0143	2.0905	0.3911 (-0.1493)	0.3501 (+0.0251)	0.4518 (-0.0488)	0.3977 (-0.0577)
4.5131	22200	2.0287	2.0905	0.4173 (-0.1231)	0.3526 (+0.0276)	0.4563 (-0.0443)	0.4087 (-0.0466)
4.5334	22300	2.0156	2.0917	0.4105 (-0.1299)	0.3520 (+0.0269)	0.4495 (-0.0512)	0.4040 (-0.0514)
4.5538	22400	2.0244	2.0905	0.4012 (-0.1393)	0.3461 (+0.0210)	0.4506 (-0.0501)	0.3993 (-0.0561)
4.5741	22500	2.0296	2.0901	0.3983 (-0.1421)	0.3502 (+0.0252)	0.4526 (-0.0480)	0.4004 (-0.0550)
4.5944	22600	2.0293	2.0906	0.4096 (-0.1308)	0.3514 (+0.0263)	0.4467 (-0.0539)	0.4026 (-0.0528)
4.6148	22700	2.0239	2.0908	0.3976 (-0.1428)	0.3490 (+0.0240)	0.4510 (-0.0496)	0.3992 (-0.0562)
4.6351	22800	2.0255	2.0911	0.3975 (-0.1429)	0.3496 (+0.0245)	0.4579 (-0.0428)	0.4017 (-0.0537)
4.6554	22900	2.0167	2.0911	0.3908 (-0.1497)	0.3470 (+0.0219)	0.4534 (-0.0473)	0.3970 (-0.0583)
4.6757	23000	2.0205	2.0908	0.4179 (-0.1225)	0.3486 (+0.0235)	0.4516 (-0.0491)	0.4060 (-0.0493)
4.6961	23100	2.0284	2.0911	0.4310 (-0.1095)	0.3488 (+0.0238)	0.4499 (-0.0507)	0.4099 (-0.0455)
4.7164	23200	2.022	2.0910	0.4199 (-0.1205)	0.3469 (+0.0219)	0.4549 (-0.0458)	0.4072 (-0.0481)
4.7367	23300	2.0176	2.0910	0.4172 (-0.1232)	0.3449 (+0.0198)	0.4549 (-0.0458)	0.4056 (-0.0497)
4.7571	23400	2.0102	2.0915	0.4054 (-0.1351)	0.3456 (+0.0206)	0.4501 (-0.0506)	0.4003 (-0.0550)
4.7774	23500	2.0097	2.0916	0.4000 (-0.1404)	0.3442 (+0.0192)	0.4442 (-0.0564)	0.3962 (-0.0592)
4.7977	23600	2.0179	2.0907	0.3971 (-0.1433)	0.3455 (+0.0205)	0.4538 (-0.0468)	0.3988 (-0.0565)
4.8181	23700	2.0244	2.0913	0.3962 (-0.1442)	0.3444 (+0.0193)	0.4537 (-0.0470)	0.3981 (-0.0573)
4.8384	23800	2.0168	2.0903	0.4028 (-0.1376)	0.3417 (+0.0167)	0.4502 (-0.0504)	0.3982 (-0.0571)
4.8587	23900	2.022	2.0910	0.3988 (-0.1416)	0.3473 (+0.0223)	0.4520 (-0.0486)	0.3994 (-0.0560)
4.8790	24000	2.0309	2.0909	0.4078 (-0.1326)	0.3445 (+0.0195)	0.4528 (-0.0478)	0.4017 (-0.0536)
4.8994	24100	2.0208	2.0912	0.3954 (-0.1451)	0.3420 (+0.0170)	0.4514 (-0.0492)	0.3963 (-0.0591)
4.9197	24200	2.0146	2.0907	0.4085 (-0.1319)	0.3421 (+0.0171)	0.4520 (-0.0486)	0.4009 (-0.0545)
4.9400	24300	2.0259	2.0907	0.4032 (-0.1372)	0.3439 (+0.0188)	0.4523 (-0.0483)	0.3998 (-0.0556)
4.9604	24400	2.0286	2.0906	0.4032 (-0.1372)	0.3449 (+0.0199)	0.4523 (-0.0483)	0.4002 (-0.0552)
4.9807	24500	2.0188	2.0910	0.4022 (-0.1382)	0.3395 (+0.0145)	0.4508 (-0.0498)	0.3975 (-0.0578)
-1	-1	-	-	0.5898 (+0.0494)	0.3760 (+0.0510)	0.6450 (+0.1443)	0.5369 (+0.0816)

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.18
• Sentence Transformers: 5.0.0
• Transformers: 4.56.0.dev0
• PyTorch: 2.7.1+cu126
• Accelerate: 1.9.0
• Datasets: 4.0.0
• Tokenizers: 0.21.4

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",
}

ListNetLoss

@inproceedings{cao2007learning,
    title={Learning to Rank: From Pairwise Approach to Listwise Approach},
    author={Cao, Zhe and Qin, Tao and Liu, Tie-Yan and Tsai, Ming-Feng and Li, Hang},
    booktitle={Proceedings of the 24th international conference on Machine learning},
    pages={129--136},
    year={2007}
}