--- tags: - sentence-transformers - sentence-similarity - feature-extraction - dense - generated_from_trainer - dataset_size:800 - loss:ContrastiveLoss base_model: Qwen/Qwen3-Embedding-4B widget: - source_sentence: "Collects the tracing data from the given parameters.\n \ \ :param request: The Flask request.\n :param response: The flask response.\n\ \ :param error: The error occurred if any.\n :param latency: The\ \ time elapsed to process the request.\n :return: The tracing data." sentences: - "def enum_to_yaml(cls: Type[T_EnumToYAML], representer: Representer, data: T_EnumToYAML)\ \ -> ruamel.yaml.nodes.ScalarNode:\n \n return representer.represent_scalar(\n\ \ f\"!{cls.__name__}\",\n f\"{str(data)}\"\n )" - "def subclasses(self, inherited=False):\n \n data = clips.data.DataObject(self._env)\n\ \n lib.EnvClassSubclasses(self._env, self._cls, data.byref, int(inherited))\n\ \n for klass in classes(self._env, data.value):\n yield klass" - "def identify(self, req, resp, resource, uri_kwargs):\n \n header\ \ = req.get_header(, False)\n auth = header.split() if header else None\n\ \n if auth is None or auth[0].lower() != :\n return None\n\n\ \ if len(auth) != 2:\n raise HTTPBadRequest(\n \ \ \"Invalid Authorization header\",\n \"The Authorization header\ \ for Token auth should be in form:\\n\"\n \"Authorization: Token\ \ \"\n )\n\n return auth[1]" - source_sentence: "A wrapper for `os.walk` that skips hidden files and directories.\n\ \n This function does not have the parameter `topdown` from\n `os.walk`:\ \ the directories must always be recursed top-down when\n using this function.\n\ \n See also\n --------\n os.walk : For a description of the parameters" sentences: - "def expand_all(self):\n \n\n def aux(item):\n self.item(item,\ \ open=True)\n children = self.get_children(item)\n for\ \ c in children:\n aux(c)\n\n children = self.get_children(\"\ \")\n for c in children:\n aux(c)" - "def create_extended_model(model, db_penalty=None, ex_penalty=None,\n \ \ tp_penalty=None, penalties=None):\n \n\n \n model_extended\ \ = model.create_metabolic_model()\n extra_compartment = model.extracellular_compartment\n\ \n compartment_ids = set(c.id for c in model.compartments)\n\n \n if\ \ len(compartment_ids) > 0:\n logger.info(\n .format(\n \ \ .join(.format(c) for c in compartment_ids)))\n db_added =\ \ add_all_database_reactions(model_extended, compartment_ids)\n else:\n \ \ logger.warning(\n \n \n )\n db_added\ \ = set()\n\n \n logger.info(\n .format(\n extra_compartment))\n\ \ ex_added = add_all_exchange_reactions(\n model_extended, extra_compartment,\ \ allow_duplicates=True)\n\n \n boundaries = model.compartment_boundaries\n\ \ if len(boundaries) > 0:\n logger.info(\n \n \ \ .format(\n .join(.format(c1, c2) for c1, c2 in boundaries)))\n\ \ tp_added = add_all_transport_reactions(\n model_extended,\ \ boundaries, allow_duplicates=True)\n else:\n logger.warning(\n \ \ \n )\n tp_added = set()\n\n \n weights = {}\n\ \ if db_penalty is not None:\n weights.update((rxnid, db_penalty) for\ \ rxnid in db_added)\n if tp_penalty is not None:\n weights.update((rxnid,\ \ tp_penalty) for rxnid in tp_added)\n if ex_penalty is not None:\n \ \ weights.update((rxnid, ex_penalty) for rxnid in ex_added)\n\n if penalties\ \ is not None:\n for rxnid, penalty in iteritems(penalties):\n \ \ weights[rxnid] = penalty\n return model_extended, weights" - "def walk_skip_hidden(top, onerror=None, followlinks=False):\n \n\n for\ \ root, dirs, files in os.walk(\n top, topdown=True, onerror=onerror,\n\ \ followlinks=followlinks):\n \n \n dirs[:] =\ \ [d for d in dirs if not is_path_hidden(d)]\n files[:] = [f for f in files\ \ if not is_path_hidden(f)]\n yield root, dirs, files" - source_sentence: Show stack frames for a task sentences: - "def do_where(self, taskid: int) -> None:\n \n task = task_by_id(taskid,\ \ self._loop)\n if task:\n self._sout.write(_format_stack(task))\n\ \ self._sout.write()\n else:\n self._sout.write(\ \ % taskid)" - "def apt_add_repository_from_apt_string(apt_string, apt_file):\n \n\n apt_file_path\ \ = % apt_file\n\n if not file_contains(apt_file_path, apt_string.lower(),\ \ use_sudo=True):\n file_append(apt_file_path, apt_string.lower(), use_sudo=True)\n\ \n with hide(, ):\n sudo(\"DEBIAN_FRONTEND=noninteractive apt-get\ \ update\")" - "def _kl_laplace_laplace(a, b, name=None):\n \n with tf.name_scope(name or \"\ kl_laplace_laplace\"):\n \n \n distance = tf.abs(a.loc - b.loc)\n \ \ ratio = a.scale / b.scale\n\n return (-tf.math.log(ratio) - 1 + distance\ \ / b.scale +\n ratio * tf.exp(-distance / a.scale))" - source_sentence: Read the ical file sentences: - "def read_ical(self, ical_file_location): \n \n with open(ical_file_location,\ \ ) as ical_file:\n data = ical_file.read()\n self.cal = Calendar.from_ical(data)\n\ \ return self.cal" - "def _size_from_header(cls, header):\n \n\n \n result = []\n\ \n for data in header:\n \n\n \n result.append(header[data])\n\ \n \n return result" - "def dockermachine_ip() -> Optional[str]:\n \n if not check_dockermachine():\n\ \ return None\n\n \n try:\n out = subprocess.check_output([,\ \ ])\n return out.decode(\"utf-8\").strip()\n except Exception:\n \ \ logger.debug(f\"docker machine not present\")\n return None" - source_sentence: 'list[VolumeExtent]: sections.' sentences: - "def cast_to_a1_notation(method):\n \n @wraps(method)\n def wrapper(self,\ \ *args, **kwargs):\n try:\n if len(args):\n \ \ int(args[0])\n\n \n range_start = rowcol_to_a1(*args[:2])\n\ \ range_end = rowcol_to_a1(*args[-2:])\n range_name = .join((range_start,\ \ range_end))\n\n args = (range_name,) + args[4:]\n except ValueError:\n\ \ pass\n\n return method(self, *args, **kwargs)\n\n return\ \ wrapper" - "def readfmt(self, fmt):\n \n size = struct.calcsize(fmt)\n \ \ blob = self.read(size)\n obj, = struct.unpack(fmt, blob)\n \ \ return obj" - "def equals(self, other):\n \n self._run(unittest_case.assertEqual,\ \ (self._subject, other))\n return ChainInspector(self._subject)" pipeline_tag: sentence-similarity library_name: sentence-transformers --- # SentenceTransformer based on Qwen/Qwen3-Embedding-4B This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Qwen/Qwen3-Embedding-4B](https://huggingface.co/Qwen/Qwen3-Embedding-4B). It maps sentences & paragraphs to a 2560-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:** [Qwen/Qwen3-Embedding-4B](https://huggingface.co/Qwen/Qwen3-Embedding-4B) - **Maximum Sequence Length:** 40960 tokens - **Output Dimensionality:** 2560 dimensions - **Similarity Function:** Cosine Similarity ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 40960, 'do_lower_case': False, 'architecture': 'Qwen3Model'}) (1): Pooling({'word_embedding_dimension': 2560, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': True, '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("TakalaWang/qwen3-embedding-4B-code-search") # Run inference queries = [ "list[VolumeExtent]: sections.", ] documents = [ 'def equals(self, other):\n \n self._run(unittest_case.assertEqual, (self._subject, other))\n return ChainInspector(self._subject)', 'def cast_to_a1_notation(method):\n \n @wraps(method)\n def wrapper(self, *args, **kwargs):\n try:\n if len(args):\n int(args[0])\n\n \n range_start = rowcol_to_a1(*args[:2])\n range_end = rowcol_to_a1(*args[-2:])\n range_name = .join((range_start, range_end))\n\n args = (range_name,) + args[4:]\n except ValueError:\n pass\n\n return method(self, *args, **kwargs)\n\n return wrapper', 'def readfmt(self, fmt):\n \n size = struct.calcsize(fmt)\n blob = self.read(size)\n obj, = struct.unpack(fmt, blob)\n return obj', ] query_embeddings = model.encode_query(queries) document_embeddings = model.encode_document(documents) print(query_embeddings.shape, document_embeddings.shape) # [1, 2560] [3, 2560] # Get the similarity scores for the embeddings similarities = model.similarity(query_embeddings, document_embeddings) print(similarities) # tensor([[1., 1., 1.]], dtype=torch.float16) ``` ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 800 training samples * Columns: sentence_0, sentence_1, and label * Approximate statistics based on the first 800 samples: | | sentence_0 | sentence_1 | label | |:--------|:------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------|:--------------------------------------------------------------| | type | string | string | float | | details | | | | * Samples: | sentence_0 | sentence_1 | label | |:---------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------| | Go through a stream and print out anything not in observed set | def t_VAR(self, t):
r
t.type = self.reserved.get(t.value.lower(), )
return t | 0.0 | | Move a page to before some other page of the document. Specify 'to = -1' to move after last page. | def movePage(self, pno, to = -1):

pl = list(range(len(self)))
if pno < 0 or pno > pl[-1]:
raise ValueError(" page number out of range")
if to < -1 or to > pl[-1]:
raise ValueError(" page number out of range")
pl.remove(pno)
if to == -1:
pl.append(pno)
else:
pl.insert(to-1, pno)
return self.select(pl) | 1.0 | | Create an empty dataset in the current repo. | def libvlc_media_player_set_agl(p_mi, drawable):

f = _Cfunctions.get(, None) or \
_Cfunction(, ((1,), (1,),), None,
None, MediaPlayer, ctypes.c_uint32)
return f(p_mi, drawable) | 0.0 | * Loss: [ContrastiveLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastiveloss) with these parameters: ```json { "distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE", "margin": 0.5, "size_average": true } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `per_device_train_batch_size`: 1 - `per_device_eval_batch_size`: 1 - `num_train_epochs`: 2 - `multi_dataset_batch_sampler`: round_robin #### All Hyperparameters
Click to expand - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: no - `prediction_loss_only`: True - `per_device_train_batch_size`: 1 - `per_device_eval_batch_size`: 1 - `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`: 5e-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`: 2 - `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`: 42 - `data_seed`: None - `jit_mode_eval`: False - `use_ipex`: False - `bf16`: False - `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`: False - `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 - `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`: round_robin - `router_mapping`: {} - `learning_rate_mapping`: {}
### Training Logs | Epoch | Step | Training Loss | |:-----:|:----:|:-------------:| | 0.625 | 500 | 0.0583 | | 1.25 | 1000 | 0.0635 | | 1.875 | 1500 | 0.0638 | ### Framework Versions - Python: 3.13.7 - Sentence Transformers: 5.1.1 - Transformers: 4.56.2 - PyTorch: 2.8.0+cu128 - Accelerate: 1.10.1 - Datasets: 4.1.1 - Tokenizers: 0.22.1 ## 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", } ``` #### ContrastiveLoss ```bibtex @inproceedings{hadsell2006dimensionality, author={Hadsell, R. and Chopra, S. and LeCun, Y.}, booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)}, title={Dimensionality Reduction by Learning an Invariant Mapping}, year={2006}, volume={2}, number={}, pages={1735-1742}, doi={10.1109/CVPR.2006.100} } ```