shivamgoel97/largefinetune

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2. It maps sentences & paragraphs to a 384-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: sentence-transformers/all-MiniLM-L6-v2
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 256, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("shivamgoel97/largefinetune")
# Run inference
sentences = [
    'What information is required when calling CMD to report an issue?',
    'Context: is due. To dismiss the alert when it is Due Now, click the MR ticket number and access the ticket.Once a scheduled follow up has arrived, notifications display as a reminder to handle the follow up.Alert: Do not place confidential, CPI or internal comments in the message box as these notes are viewable by the customer. Statusing Follow Ups Follow Up What If ScenariosManages Resolution (MR) Follow Up (automated) enhancements give customers more information and control during the process. After a ticket is submitted, review below to see what the customer sees and how the customer can manage their ticket.Table: <table border="1" cellpadding="5" cellspacing="0" width="100%"><tbody><tr><td style="text-align: left;" valign="top" width="35%"><b>If ...</b></td><td style="text-align: left;" valign="top" width="65%"><b>Then ...</b></td></tr><tr><td style="text-align: left;" valign="top" width="35%">agent misses the scheduled window for contact</td><td style="text-align: left;" valign="top" width="65%">agent attempts to contact. If customer does not answer, request to reschedule and do so when customer responds.\xa0</td></tr><tr><td style="text-align: left;" valign="top" width="35%">the agent is out sick or on vacation</td><td style="text-align: left;" valign="top" width="65%">supervisors reassign the follow up to ensure the customer receives the call back.</td></tr><tr><td style="text-align: left;" valign="top" width="35%">the customer modifies the scheduled call back</td><td style="text-align: left;" valign="top" width="65%">customer selects a new time based on the agent\'s availability through My Verizon app. The follow up system automatically updates the time on the agent\'s end and provides reminders the day of the newly selected time.\xa0</td></tr><tr><td style="text-align: left;" valign="top" width="35%">the customer cannot be reached</td><td style="text-align: left;" valign="top" width="65%"><p>the agent completes the following:</p><ol><li>Leave a detailed voicemail.</li><li>Extend the follow up for the next business day.\xa0</li><li>Select a reason for the reschedule:<ul><li>No answer</li><li>Voicemail</li><li>Line busy</li><li>Note: Selecting 1 of these options places the ticket in Pending-Need more information status.\xa0</li></ul></li><li>Update the Follow Up Status accordingly.</li></ol></td></tr><tr><td style="text-align: left;" valign="top" width="35%">there is a need to make changes to a follow up without agreement from the customer</td><td style="text-align: left;" valign="top" width="65%"><p>do not do without customer approval.\xa0</p><p>Note: The Managed Resolution (MR) Follow Up transformation offers customers the experience to receive updates on their follow ups and receive updates any time there is a change. Customer may call back if it is not a change the customer requested (e.g., changing MDN, follow up method, etc.)</p></td></tr><tr><td style="text-align: left;" valign="top">there is a need to make a change to a follow up and waiting for the customer to respond</td><td style="text-align: left;" valign="top"><p>select <b>Need more information</b>.\xa0</p><p>Result: The customer receives a 24 hour and 72 hour reminder. If the customer does not respond within 7 days, the follow up auto-resolves.</p></td></tr><tr><td style="text-align: left;" valign="top">waiting for NRB ticket or another resolution prior to contacting the customer</td><td style="text-align: left;" valign="top"><p>select <b>Need more time, still reviewing and processing</b>.</p><p>Agent types an update message to the customer that is sent through SMS or email to alert the customer of the follow up status.</p></td></tr></tbody></table>',
    "Assist customers who are experiencing trouble adding the Verizon Connections discount to their account. This enhancement is for Verizon Connections and does not include the Veterans Advantage Affinity Program. Add the token to New Installs, Moves, Change, or Pending orders. Adding VZ Connection Token in Optix Select VZ Connections in the dropdown in the Coupons, Tokens and Discounts section of One Page Ordering (OPO). Enter the 12-digit alphanumeric token provided by the customer. The token is specific to the customer and can only be provided by them. It is not in Optix. Click Verify Customer Coupon to validate the token. Click Apply to add the discount to the customer's account. The Verizon Connections discount is presented on the Agreement + Offer section tile: Click Review Order. The discount appears on the Review Order page:",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Training Details

Training Dataset

Unnamed Dataset

• Size: 30,000 training samples
• Columns: sentence_0, sentence_1, and label
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	label
  type	string	string	int
  details	min: 9 tokens mean: 18.77 tokens max: 35 tokens	min: 55 tokens mean: 218.68 tokens max: 256 tokens	0: ~32.80% 1: ~67.20%

• Samples:

  sentence_0	sentence_1	label
  Who is exempt from the rules set by the Truth in Caller ID Act?	The Truth in Caller ID Act The FCC adopted rules implementing the Truth in Caller ID Act. The FCC rules: Prohibit any person or entity from transmitting misleading or inaccurate caller ID information with the intent to defraud, cause harm or wrongfully get anything of value. Subject violators to a penalty of up to $10,000 for each violation of the rules. Exempt authorized activities by law enforcement agencies and situations where courts have authorized caller ID manipulation to occur. Need additional assistance with walking the customer through their experience? Access the IG Fraud Response tool for helpful scenarios and responses. Verizon wireless also offers Caller Name ID; which can be downloaded on user's device through the APP Store (preloaded on Android devices). Caller ID has Spam Management functionality. See Call Filter for additional details (monthly fee applies). Refer the customer to the FCC Spoofing and Caller ID website for additional information. There is a short video ...	1
  What additional tools does the paid version of the call filtering feature provide?	Sell it in 5: Call Filter (free) helps companies and agencies increase productivity and take control of their calls by identifying and routing any spam, fraudulent, or illegal robocalls to voicemail. Call Filter Plus (paid version) provides additional tools for user control such as manually blocking and unblocking numbers; blocking calls based on risk levels; and adding a name to unknown numbers in order to minimize impact to a business or agency. Sales Point of Contacts (SPOC's) may order or block the free and paid versions of the feature for their employees by customizing the feature block on their applicable lines. Business employees may order Call Filter (free) and engage their company SPOC to order Call Filter Plus (paid version through contract amendment). Public Sector employees engage their SPOC to order Call Filter or Call Filter Plus through a contract amendment.	1
  What features are included with the Unlimited storage option that are not available with the 600 GB option?	Context: Connection 2.0 will not receive this trial period since there is no cost. Customer communications and setup flows will not mention the "30 days on us" when 2TB is included in Gigabit Connection 2.0. Current In-Market 500 GB: $5/month (available for prepaid customers only) 600 GB: $5.99/month Effective 9/15/20, the 500 GB Verizon Cloud storage option was increased to 600 GB for $5.99. Customers who had the 500 GB storage option accounts were automatically increased to 600 GB of storage for $5.99/month. A notification about this change was sent to the account owner through their Verizon account. 1 TB: $9.99/monthTable: Offers as of 12/2/20      Unlimited $19.99/month 2 TB $14.99/month 600 GB $5.99/month Users and devices Up to 5 users Unlimited devices Up to 5 users Unlimited devices 1 Loss: CosineSimilarityLoss with these parameters: { "loss_fct": "torch.nn.modules.loss.MSELoss" } Training Hyperparameters Non-Default Hyperparameters 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: 8 per_device_eval_batch_size: 8 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 num_train_epochs: 3 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} 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 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 dispatch_batches: None split_batches: 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 eval_use_gather_object: False average_tokens_across_devices: False prompts: None batch_sampler: batch_sampler multi_dataset_batch_sampler: round_robin Training Logs Epoch Step Training Loss 0.1333 500 0.179 0.2667 1000 0.1591 0.4 1500 0.1492 0.5333 2000 0.1454 0.6667 2500 0.1422 0.8 3000 0.1404 0.9333 3500 0.1354 1.0667 4000 0.1288 1.2 4500 0.1256 1.3333 5000 0.1234 1.4667 5500 0.1208 1.6 6000 0.1215 1.7333 6500 0.123 1.8667 7000 0.1231 2.0 7500 0.1181 2.1333 8000 0.1061 2.2667 8500 0.1066 2.4 9000 0.1093 2.5333 9500 0.105 2.6667 10000 0.1088 2.8 10500 0.1053 2.9333 11000 0.1062 Framework Versions Python: 3.10.12 Sentence Transformers: 3.3.1 Transformers: 4.47.1 PyTorch: 2.5.1+cu121 Accelerate: 1.2.1 Datasets: 3.2.0 Tokenizers: 0.21.0 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", }