Modelpublic

HariishHafiiz/sbert-bug-eclipse-ft

sourceHugging Faceupdated 4mo agoView on Hugging Face

0likes87downloads

Model Card

SentenceTransformer based on sentence-transformers/all-mpnet-base-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-mpnet-base-v2. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for retrieval.

Model Details

Model Description

—Model Type: Sentence Transformer
—Base model: sentence-transformers/all-mpnet-base-v2
—Maximum Sequence Length: 256 tokens
—Output Dimensionality: 768 dimensions
—Similarity Function: Cosine Similarity
—Supported Modality: Text

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': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'MPNetModel'})
  (1): Pooling({'embedding_dimension': 768, 'pooling_mode': 'mean', 'include_prompt': True})
  (2): Normalize({})
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

bash

pip install -U sentence-transformers

Then you can load this model and run inference.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
    'tooltips and table columns please add tooltips to the swt table columns.  i would like to give each  column a short description that popsup when a user hovers over it.    for a large table editor it would be helpful to have a quick reminder of what  the column is for.  hovering over the column header could provide that  information without making the user go into the online help.',
    'need support for tool tips in table columns gbzld table column headings should support tool tips.  trigger pr gqkt itpjuiwinnt  tasklist columns  no tooltip and bogus names     nepm \tplease advise when status of this pr changes.  \tmcqpm  \t\twe will not have time to get to this.resolvedfixed.',
    'logj.properties file not found i have a java project. the java project uses logj. i have a logj.properties  file in the same directory as the ant .xml file. i have in a separate  source directory but the same package a set of junit tests.  i go to my command prompt. i go to the directory with the .xml file and  say  ant utest  and all my tests run and pass. now i go to eclipse. i can run all my targets    with unit tests create dist pack create jar export jar to  other projects javadoc etc etc... except one. the one that does not run  is utest. the error is  utest logjerror could not read configuration file logj.properties. java.io.filenotfoundexception logj.properties the system cannot find the file specified  . i have set the working directory on the run ant dialog to the correct  place  no difference. . i have searched your bugs for logj  no results.  thus... at a guess it seems like eclipse is losing my working directory and  hence logj.properties is not found.  i include the utest target  project defaultu basedir.  target nameutest dependsu \tjunit \t\tclasspath refidproject.classpath \t\ttest nametest \t\tformatter typebrief usefileno \tjunit target  project  where u works just fine and the test name is alltests. this is standard  stuff and works on the command line in linux and windows and in eclipse . yes the one with the nice ant interface  which is still being  used by the rest of the team. changing the project basedir in the ant file  does not have any effect. the ant version i installed manually is the same as  the ant version that comes with eclipse . .  easiest way for you to test may be to get some old code that also has unit  tests slap in an initialisation line for logj  private static category cat  category.getinstance myclass.class   slap a logj.properties file in the same place the the ant  file for  example  logj.appender.consoleorg.apache.logj.consoleappender logj.appender.console.targetsystem.out   logj.appender.console.layoutorg.apache.logj.patternlayout logj.appender.console.layout.conversionpatternp t c m line  ln logj.rootcategoryfatalconsole  add the utest target to your  file and then try it with an older and a  new version of eclipse.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[ 1.0000,  0.7714, -0.0956],
#         [ 0.7714,  1.0000, -0.0372],
#         [-0.0956, -0.0372,  1.0000]])

<!--

Direct Usage (Transformers)

<details><summary>Click to see the direct usage in Transformers</summary>

</details> -->

<!--

Downstream Usage (Sentence Transformers)

You can finetune this model on your own dataset.

<details><summary>Click to expand</summary>

</details> -->

<!--

Out-of-Scope Use

List how the model may foreseeably be misused and address what users ought not to do with the model. -->

<!--

Bias, Risks and Limitations

What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. -->

<!--

Recommendations

What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. -->

Training Details

Training Dataset

Unnamed Dataset

—Size: 5,920 training samples
—Columns: <code>anchor</code> and <code>positive</code>
—Approximate statistics based on the first 1000 samples: | | anchor | positive | |:--------|:-------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------| | type | string | string | | details | <ul><li>min: 17 tokens</li><li>mean: 129.86 tokens</li><li>max: 256 tokens</li></ul> | <ul><li>min: 13 tokens</li><li>mean: 130.06 tokens</li><li>max: 256 tokens</li></ul> |
—Samples: | anchor | positive | |:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | <code>ui viewseditors appear active when inactive getkg if you switch to a process outside of eclipse the native active title bar becomes inactive and turns grey. the eclipse views and editors however remain coloured which may appear confusing. to repeat select a view or editor in eclipse note the title bar colourize. switch to another app and restore down so that you can see eclipse in the background.</code> | <code>confusing inner windows active when outer inactive gjnku when looking at an inactive eclipse workbench the title is grey to indicate that it is not active but the inner windows viewparts remain in the active color usually the gradiant blue. this is confusing since a cursory look at the window indicates that it is active and then after trying to do something like pressing f to continue to debug it does not work.</code> | | <code>cvs ui creating a patch after creating a new folder does not work properly create a new project get it synced with a cvs repo. create a new folder containing a file in the project root but dont add either of them to cvs. ceate a patch for the project unified output format and save it to either the clipbard or the filesystem. now delete the new folder and apply the patch. it creates only the file but puts it into the project root instead of creating the folder and placing the file there.</code> | <code>cvs patch create patch misses new directories i apologize if someone has already reported this i couldnt find a duplicate. if you create a new patch and choose include new files in patch eclipse very unintuitively ignores new files which exist in new directories. please either fix the wording to indicate that new directories will be ignored or preferably include the new files in the patch.</code> | | <code>changing the properties of a cvs location blocks the ide i tried to change the properties of a cvs location. after pressing the ok or apply button a modal dialog progress information appears with buttons details and cancel disabled. behind this dialog another dialog appears which asks me to confirm project sharing changes and blocks the progress. the second dialog cannot be reached because it is blocked by the first dialog. the only way out seems to kill the eclipse process and lose all unsaved changes. i only tried this on mac os x . but maybe the bug occurs also on other operating systems.</code> | <code>cant change repository parameters open the cvs repository exploring perspective. select an existing pserver repository that has attached projects. click properties. change the connection type to extssh and click either apply or ok. a dialog box with a title bar progress information opens. inside that box it says operation in progress... behind that dialog another opens that says confirm project sharing changes. there are options ok details cancel. since it is in the background behind the progress information box you cant click any option. while you can move the progress information box out of the way you cant make the confirm project sharing changes box the foreground window. you can no longer click to any other window but must quit eclipse to make it workable again.</code> |
—Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

json

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }

Evaluation Dataset

Unnamed Dataset

—Size: 1,480 evaluation samples
—Columns: <code>anchor</code> and <code>positive</code>
—Approximate statistics based on the first 1000 samples: | | anchor | positive | |:--------|:-------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------| | type | string | string | | details | <ul><li>min: 18 tokens</li><li>mean: 134.54 tokens</li><li>max: 256 tokens</li></ul> | <ul><li>min: 18 tokens</li><li>mean: 130.34 tokens</li><li>max: 256 tokens</li></ul> |
—Samples: | anchor | positive | |:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | <code>ant editor auto complete missing ftp task the auto complete feature in the ant editor does not show the ftp task. i use eclipse .. and eclipse .m s. i run windows xp home editon at home and windows pro at work. both systems and version of eclipse have this problem. just try to create an ant script and try the auto complete feature.</code> | <code>update the ant code assist support files for ant .. the support files for the ant code assist need to be updated for ant .. the files are ant.b.dtd and anttasks.b.xml</code> | | <code>disabling breakpoint causes focus to move if you have a laundry list of breakpoints i.e. scroll bar appear in the breakpoints view when you disable one the focus jumps to the item that is selected. its awkward when you are trying to disable a set of breakpoints at the bottom when the select is at the top.</code> | <code>disabling breakpoint causes focus to move if you have a laundry list of breakpoints i.e. scroll bar appear in the breakpoints view when you disable one the focus jumps to the item that is selected. its awkward when you are trying to disable a set of breakpoints at the bottom when the select is at the top.</code> | | <code>swt doesnt allow client code to change the drag cursor the drag cursors on swt are set to the default system drag cursors. applications that want to be richer cannot modify this eventhough it is possible in mostall the platforms that matter.</code> | <code>dcr support for changing cursor in dnd i have a problem with the dndfeedback in the package viewer whenver i dnd something over a nonacceptable drop target i get a mouse cursor which indicates that i cannot drop my load circle but i dont see a mouse cursor with a clearly defined hotspot and some indication that im carrying a load e.g. the arrow with smal dotted rectangle. i would suggest to use the arrow with box cursor for the full duration of the dnd operation and give the acceptance feedback either on the target or if this isnt possible as a small modification of the arrow with box. this would ensure that the user knows where the mouse cursor points to and that he is performing a dnd operation. db .. i think this is platform behaviour but moving to swt to make sure. cmpm vi to investigate. viam currently we are using the platform default cursors for drag and drop. the do not enter circle is standard windows behaviour. using the arrow with box cursor would im...</code> |
—Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

json

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }

Training Hyperparameters

Non-Default Hyperparameters

—eval_strategy: steps
—weight_decay: 0.01
—num_train_epochs: 1
—warmup_steps: 100
—fp16: True

All Hyperparameters

<details><summary>Click to expand</summary>

—do_predict: False
—eval_strategy: steps
—prediction_loss_only: True
—per_device_train_batch_size: 8
—per_device_eval_batch_size: 8
—gradient_accumulation_steps: 1
—eval_accumulation_steps: None
—torch_empty_cache_steps: None
—learning_rate: 5e-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: 1
—max_steps: -1
—lr_scheduler_type: linear
—lr_scheduler_kwargs: None
—warmup_ratio: None
—warmup_steps: 100
—log_level: passive
—log_level_replica: warning
—log_on_each_node: True
—logging_nan_inf_filter: True
—enable_jit_checkpoint: False
—save_on_each_node: False
—save_only_model: False
—restore_callback_states_from_checkpoint: False
—use_cpu: False
—seed: 42
—data_seed: None
—bf16: False
—fp16: True
—bf16_full_eval: False
—fp16_full_eval: False
—tf32: None
—local_rank: -1
—ddp_backend: None
—debug: []
—dataloader_drop_last: False
—dataloader_num_workers: 0
—dataloader_prefetch_factor: None
—disable_tqdm: False
—remove_unused_columns: True
—label_names: None
—load_best_model_at_end: False
—ignore_data_skip: False
—fsdp: []
—fsdp_config: {'minnumparams': 0, 'xla': False, 'xlafsdpv2': False, 'xlafsdpgrad_ckpt': False}
—accelerator_config: {'splitbatches': False, 'dispatchbatches': None, 'evenbatches': True, 'useseedablesampler': True, 'nonblocking': False, 'gradientaccumulationkwargs': None}
—parallelism_config: None
—deepspeed: None
—label_smoothing_factor: 0.0
—optim: adamwtorchfused
—optim_args: None
—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
—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
—auto_find_batch_size: False
—full_determinism: False
—ddp_timeout: 1800
—torch_compile: False
—torch_compile_backend: None
—torch_compile_mode: None
—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
—use_cache: False
—prompts: None
—batch_sampler: batch_sampler
—multi_dataset_batch_sampler: proportional
—router_mapping: {}
—learning_rate_mapping: {}

</details>

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.0676	25	0.3345	0.2836
0.1351	50	0.2917	0.2626
0.2027	75	0.2348	0.2747
0.2703	100	0.2279	0.2513
0.3378	125	0.3163	0.2360
0.4054	150	0.2820	0.2714
0.4730	175	0.2091	0.2497
0.5405	200	0.2357	0.2328
0.6081	225	0.2256	0.2234
0.6757	250	0.2589	0.2151
0.7432	275	0.1895	0.2104
0.8108	300	0.2007	0.2176
0.8784	325	0.2382	0.2166
0.9459	350	0.2020	0.2147

Training Time

—Training: 13.7 minutes

Framework Versions

—Python: 3.12.13
—Sentence Transformers: 5.4.0
—Transformers: 5.0.0
—PyTorch: 2.10.0+cu128
—Accelerate: 1.13.0
—Datasets: 4.8.5
—Tokenizers: 0.22.2

Citation

BibTeX

Sentence Transformers

bibtex

@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

bibtex

@misc{oord2019representationlearningcontrastivepredictive,
      title={Representation Learning with Contrastive Predictive Coding},
      author={Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      year={2019},
      eprint={1807.03748},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/1807.03748},
}

<!--

Glossary

Clearly define terms in order to be accessible across audiences. -->

<!--

Model Card Authors

Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. -->

<!--

Model Card Contact

Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->