You can deploy this model and start earning money today!

KayaTechAI/
Qwen3-0.6B-Fine-Tuned-Telecom-Technical-Documents-Retrieval-Embedding-With-Config

Name: Qwen3-0.6B-Fine-Tuned-Telecom-Technical-Documents-Retrieval-Embedding-With-Config
Availability: InStock

Views

No views yet

SentenceTransformer based on Qwen/Qwen3-Embedding-0.6B

This is a sentence-transformers model finetuned from Qwen/Qwen3-Embedding-0.6B on the telecom-technical-documents-retrieval-embedding-dataset dataset. It maps sentences & paragraphs to a 1024-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-0.6B
Maximum Sequence Length: 32768 tokens
Output Dimensionality: 1024 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- telecom-technical-documents-retrieval-embedding-dataset

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': 32768, 'do_lower_case': False, 'architecture': 'Qwen3Model'})
  (1): Pooling({'word_embedding_dimension': 1024, '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:

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("KayaTechAI/Qwen3-0.6B-Fine-Tuned-Telecom-Technical-Documents-Retrieval-Embedding-With-Config")
# Run inference
queries = [
    "What is the provisioning scope for the eMLPP service?",
]
documents = [
    'eMLPP is provisioned per subscriber.',
    'The main objective is to verify that the User Equipment (UE) tracks channel variations and selects the optimal transport format for frequency non-selective scheduling.',
    'SDP is used in SIP communications to describe the parameters and media capabilities of a session, such as audio/video codecs, transport protocols, and IP addresses, enabling participants to agree on the media types to be used.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 1024] [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[ 0.6303, -0.0008, -0.0340]])

Evaluation

Metrics

Information Retrieval

Dataset: dim_1024
Evaluated with InformationRetrievalEvaluator with these parameters:
json
1{ 2 "truncate_dim": 1024 3}

Metric	Value
cosine_accuracy@1	0.7988
cosine_accuracy@3	0.912
cosine_accuracy@5	0.9404
cosine_accuracy@10	0.9636
cosine_precision@1	0.7988
cosine_precision@3	0.304
cosine_precision@5	0.1881
cosine_precision@10	0.0964
cosine_recall@1	0.7988
cosine_recall@3	0.912
cosine_recall@5	0.9404
cosine_recall@10	0.9636
cosine_ndcg@10	0.886
cosine_mrr@10	0.8606
cosine_map@100	0.8621

Information Retrieval

Dataset: dim_768
Evaluated with InformationRetrievalEvaluator with these parameters:
json
1{ 2 "truncate_dim": 768 3}

Metric	Value
cosine_accuracy@1	0.7996
cosine_accuracy@3	0.9148
cosine_accuracy@5	0.9408
cosine_accuracy@10	0.9624
cosine_precision@1	0.7996
cosine_precision@3	0.3049
cosine_precision@5	0.1882
cosine_precision@10	0.0962
cosine_recall@1	0.7996
cosine_recall@3	0.9148
cosine_recall@5	0.9408
cosine_recall@10	0.9624
cosine_ndcg@10	0.8859
cosine_mrr@10	0.8608
cosine_map@100	0.8625

Information Retrieval

Dataset: dim_512
Evaluated with InformationRetrievalEvaluator with these parameters:
json
1{ 2 "truncate_dim": 512 3}

Metric	Value
cosine_accuracy@1	0.7968
cosine_accuracy@3	0.9128
cosine_accuracy@5	0.9388
cosine_accuracy@10	0.962
cosine_precision@1	0.7968
cosine_precision@3	0.3043
cosine_precision@5	0.1878
cosine_precision@10	0.0962
cosine_recall@1	0.7968
cosine_recall@3	0.9128
cosine_recall@5	0.9388
cosine_recall@10	0.962
cosine_ndcg@10	0.8844
cosine_mrr@10	0.8589
cosine_map@100	0.8606

Information Retrieval

Dataset: dim_256
Evaluated with InformationRetrievalEvaluator with these parameters:
json
1{ 2 "truncate_dim": 256 3}

Metric	Value
cosine_accuracy@1	0.7804
cosine_accuracy@3	0.912
cosine_accuracy@5	0.9316
cosine_accuracy@10	0.9584
cosine_precision@1	0.7804
cosine_precision@3	0.304
cosine_precision@5	0.1863
cosine_precision@10	0.0958
cosine_recall@1	0.7804
cosine_recall@3	0.912
cosine_recall@5	0.9316
cosine_recall@10	0.9584
cosine_ndcg@10	0.8753
cosine_mrr@10	0.848
cosine_map@100	0.8496

Information Retrieval

Dataset: dim_128
Evaluated with InformationRetrievalEvaluator with these parameters:
json
1{ 2 "truncate_dim": 128 3}

Metric	Value
cosine_accuracy@1	0.7696
cosine_accuracy@3	0.898
cosine_accuracy@5	0.9268
cosine_accuracy@10	0.9524
cosine_precision@1	0.7696
cosine_precision@3	0.2993
cosine_precision@5	0.1854
cosine_precision@10	0.0952
cosine_recall@1	0.7696
cosine_recall@3	0.898
cosine_recall@5	0.9268
cosine_recall@10	0.9524
cosine_ndcg@10	0.8663
cosine_mrr@10	0.8381
cosine_map@100	0.8399

Information Retrieval

Dataset: dim_64
Evaluated with InformationRetrievalEvaluator with these parameters:
json
1{ 2 "truncate_dim": 64 3}

Metric	Value
cosine_accuracy@1	0.75
cosine_accuracy@3	0.8816
cosine_accuracy@5	0.9124
cosine_accuracy@10	0.9456
cosine_precision@1	0.75
cosine_precision@3	0.2939
cosine_precision@5	0.1825
cosine_precision@10	0.0946
cosine_recall@1	0.75
cosine_recall@3	0.8816
cosine_recall@5	0.9124
cosine_recall@10	0.9456
cosine_ndcg@10	0.8522
cosine_mrr@10	0.8218
cosine_map@100	0.8236

Training Details

Training Dataset

telecom-technical-documents-retrieval-embedding-dataset

Dataset: telecom-technical-documents-retrieval-embedding-dataset at 3ebf34a
Size: 127,731 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 7 tokens
mean: 18.79 tokens
max: 68 tokens
min: 4 tokens
mean: 26.09 tokens
max: 77 tokens

Samples:

anchor	positive
`What is the estimated Transmit power considered sufficient for achieving 95% Downlink coverage with a single Base Station?`	`Approximately 14 dBm Transmit power is considered sufficient.`
`What is the primary goal of the Nominal Accuracy requirement?`	`The primary goal of the Nominal Accuracy requirement is to ensure good accuracy when signal conditions are ideal.`
`What happens on the mobile station side if contention resolution fails because the G-RNTI value in the network's acknowledgement message differs from what the mobile station sent?`	`If the mobile station receives a PACKET UPLINK ACK/NACK message with a G-RNTI value different from the one it included in its first RLC data blocks, it signifies a contention resolution failure, and the mobile station will not transmit a PACKET CONTROL ACKNOWLEDGEMENT.`

Loss: MatryoshkaLoss with these parameters:
json
1{ 2 "loss": "MultipleNegativesRankingLoss", 3 "matryoshka_dims": [ 4 1024, 5 768, 6 512, 7 256, 8 128, 9 64 10 ], 11 "matryoshka_weights": [ 12 1, 13 1, 14 1, 15 1, 16 1, 17 1 18 ], 19 "n_dims_per_step": -1 20}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
gradient_accumulation_steps: 16
learning_rate: 2e-05
num_train_epochs: 4
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: True
tf32: True
load_best_model_at_end: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 16
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: 4
max_steps: -1
lr_scheduler_type: cosine
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: True
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_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: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	dim_1024_cosine_ndcg@10	dim_768_cosine_ndcg@10	dim_512_cosine_ndcg@10	dim_256_cosine_ndcg@10	dim_128_cosine_ndcg@10	dim_64_cosine_ndcg@10
0.0401	10	1.5256	-	-	-	-	-	-
0.0802	20	0.8247	-	-	-	-	-	-
0.1202	30	0.4102	-	-	-	-	-	-
0.1603	40	0.27	-	-	-	-	-	-
0.2004	50	0.2182	-	-	-	-	-	-
0.2405	60	0.1998	-	-	-	-	-	-
0.2806	70	0.2017	-	-	-	-	-	-
0.3206	80	0.1672	-	-	-	-	-	-
0.3607	90	0.2029	-	-	-	-	-	-
0.4008	100	0.1609	-	-	-	-	-	-
0.4409	110	0.1565	-	-	-	-	-	-
0.4810	120	0.1476	-	-	-	-	-	-
0.5210	130	0.1278	-	-	-	-	-	-
0.5611	140	0.1669	-	-	-	-	-	-
0.6012	150	0.1642	-	-	-	-	-	-
0.6413	160	0.1307	-	-	-	-	-	-
0.6814	170	0.1487	-	-	-	-	-	-
0.7214	180	0.1329	-	-	-	-	-	-
0.7615	190	0.13	-	-	-	-	-	-
0.8016	200	0.1393	-	-	-	-	-	-
0.8417	210	0.1344	-	-	-	-	-	-
0.8818	220	0.1184	-	-	-	-	-	-
0.9218	230	0.1147	-	-	-	-	-	-
0.9619	240	0.1283	-	-	-	-	-	-
1.0	250	0.1228	0.8693	0.8683	0.8634	0.8535	0.8430	0.8082
1.0401	260	0.0613	-	-	-	-	-	-
1.0802	270	0.0559	-	-	-	-	-	-
1.1202	280	0.0704	-	-	-	-	-	-
1.1603	290	0.0578	-	-	-	-	-	-
1.2004	300	0.0588	-	-	-	-	-	-
1.2405	310	0.079	-	-	-	-	-	-
1.2806	320	0.0602	-	-	-	-	-	-
1.3206	330	0.0553	-	-	-	-	-	-
1.3607	340	0.0663	-	-	-	-	-	-
1.4008	350	0.0513	-	-	-	-	-	-
1.4409	360	0.0615	-	-	-	-	-	-
1.4810	370	0.0462	-	-	-	-	-	-
1.5210	380	0.0674	-	-	-	-	-	-
1.5611	390	0.0558	-	-	-	-	-	-
1.6012	400	0.0562	-	-	-	-	-	-
1.6413	410	0.0688	-	-	-	-	-	-
1.6814	420	0.0905	-	-	-	-	-	-
1.7214	430	0.0463	-	-	-	-	-	-
1.7615	440	0.0581	-	-	-	-	-	-
1.8016	450	0.0586	-	-	-	-	-	-
1.8417	460	0.0712	-	-	-	-	-	-
1.8818	470	0.041	-	-	-	-	-	-
1.9218	480	0.0578	-	-	-	-	-	-
1.9619	490	0.063	-	-	-	-	-	-
2.0	500	0.0505	0.8771	0.8780	0.8764	0.8690	0.8587	0.8353
2.0401	510	0.032	-	-	-	-	-	-
2.0802	520	0.0239	-	-	-	-	-	-
2.1202	530	0.029	-	-	-	-	-	-
2.1603	540	0.0236	-	-	-	-	-	-
2.2004	550	0.0381	-	-	-	-	-	-
2.2405	560	0.028	-	-	-	-	-	-
2.2806	570	0.0366	-	-	-	-	-	-
2.3206	580	0.0372	-	-	-	-	-	-
2.3607	590	0.0306	-	-	-	-	-	-
2.4008	600	0.0294	-	-	-	-	-	-
2.4409	610	0.0269	-	-	-	-	-	-
2.4810	620	0.0411	-	-	-	-	-	-
2.5210	630	0.0251	-	-	-	-	-	-
2.5611	640	0.0299	-	-	-	-	-	-
2.6012	650	0.0275	-	-	-	-	-	-
2.6413	660	0.0267	-	-	-	-	-	-
2.6814	670	0.0304	-	-	-	-	-	-
2.7214	680	0.0246	-	-	-	-	-	-
2.7615	690	0.025	-	-	-	-	-	-
2.8016	700	0.037	-	-	-	-	-	-
2.8417	710	0.0393	-	-	-	-	-	-
2.8818	720	0.0405	-	-	-	-	-	-
2.9218	730	0.0279	-	-	-	-	-	-
2.9619	740	0.0243	-	-	-	-	-	-
3.0	750	0.0284	0.8870	0.8858	0.8827	0.8745	0.8648	0.8499
3.0401	760	0.0166	-	-	-	-	-	-
3.0802	770	0.024	-	-	-	-	-	-
3.1202	780	0.0302	-	-	-	-	-	-
3.1603	790	0.0263	-	-	-	-	-	-
3.2004	800	0.0172	-	-	-	-	-	-
3.2405	810	0.023	-	-	-	-	-	-
3.2806	820	0.0313	-	-	-	-	-	-
3.3206	830	0.0253	-	-	-	-	-	-
3.3607	840	0.0189	-	-	-	-	-	-
3.4008	850	0.0177	-	-	-	-	-	-
3.4409	860	0.0187	-	-	-	-	-	-
3.4810	870	0.0142	-	-	-	-	-	-
3.5210	880	0.0281	-	-	-	-	-	-
3.5611	890	0.0253	-	-	-	-	-	-
3.6012	900	0.0184	-	-	-	-	-	-
3.6413	910	0.0217	-	-	-	-	-	-
3.6814	920	0.027	-	-	-	-	-	-
3.7214	930	0.0192	-	-	-	-	-	-
3.7615	940	0.0183	-	-	-	-	-	-
3.8016	950	0.0242	-	-	-	-	-	-
3.8417	960	0.0223	-	-	-	-	-	-
3.8818	970	0.0161	-	-	-	-	-	-
3.9218	980	0.0219	-	-	-	-	-	-
3.9619	990	0.0236	-	-	-	-	-	-
4.0	1000	0.0278	0.886	0.8859	0.8844	0.8753	0.8663	0.8522

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.12.12
Sentence Transformers: 5.2.3
Transformers: 4.55.4
PyTorch: 2.10.0+cu128
Accelerate: 1.12.0
Datasets: 3.6.0
Tokenizers: 0.21.4

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",
}

MatryoshkaLoss

bibtex

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

MultipleNegativesRankingLoss

bibtex

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

	anchor	positive
type	string	string
details	min: 7 tokens mean: 18.79 tokens max: 68 tokens	min: 4 tokens mean: 26.09 tokens max: 77 tokens