SentenceTransformer based on Qwen/Qwen2.5-0.5B-Instruct

This is a sentence-transformers model finetuned from Qwen/Qwen2.5-0.5B-Instruct. It maps sentences & paragraphs to a 896-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/Qwen2.5-0.5B-Instruct
Maximum Sequence Length: 1024 tokens
Output Dimensionality: 896 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': 1024, 'do_lower_case': False}) with Transformer model: Qwen2Model 
  (1): Pooling({'word_embedding_dimension': 896, '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})
)

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("AlexWortega/qwen3k")
# Run inference
sentences = [
    'When was ABC formed?',
    "American Broadcasting Company\nABC launched as a radio network on October 12, 1943, serving as the successor to the NBC Blue Network, which had been purchased by Edward J. Noble. It extended its operations to television in 1948, following in the footsteps of established broadcast networks CBS and NBC. In the mid-1950s, ABC merged with United Paramount Theatres, a chain of movie theaters that formerly operated as a subsidiary of Paramount Pictures. Leonard Goldenson, who had been the head of UPT, made the new television network profitable by helping develop and greenlight many successful series. In the 1980s, after purchasing an 80% interest in cable sports channel ESPN, the network's corporate parent, American Broadcasting Companies, Inc., merged with Capital Cities Communications, owner of several print publications, and television and radio stations. In 1996, most of Capital Cities/ABC's assets were purchased by The Walt Disney Company.",
    'Americans Battling Communism\nAmericans Battling Communism, Inc. (ABC) was an anti-communist organization created following an October 1947 speech by Pennsylvania Judge Blair Gunther that called for an "ABC movement" to educate America about communism. Chartered in November 1947 by Harry Alan Sherman, a local lawyer active in various anti-communist organizations, the group took part in such activities as blacklisting by disclosing the names of people suspected of being communists. Its members included local judges and lawyers active in the McCarthy-era prosecution of communists.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 896]

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

Evaluation

Metrics

Semantic Similarity

Datasets: sts-dev-896 and sts-dev-768
Evaluated with EmbeddingSimilarityEvaluator

Metric	sts-dev-896	sts-dev-768
pearson_cosine	0.7513	0.7504
spearman_cosine	0.7603	0.759

Training Details

Training Dataset

Unnamed Dataset

Size: 1,077,240 training samples
Columns: query, response, and negative

Approximate statistics based on the first 1000 samples:

	query	response	negative
type	string	string	string
details	min: 4 tokens mean: 8.76 tokens max: 26 tokens	min: 23 tokens mean: 141.88 tokens max: 532 tokens	min: 4 tokens mean: 134.02 tokens max: 472 tokens

Samples:

query	response	negative
`Was there a year 0?`	`Year zero Year zero does not exist in the anno Domini system usually used to number years in the Gregorian calendar and in its predecessor, the Julian calendar. In this system, the year 1 BC is followed by AD 1. However, there is a year zero in astronomical year numbering (where it coincides with the Julian year 1 BC) and in ISO 8601:2004 (where it coincides with the Gregorian year 1 BC) as well as in all Buddhist and Hindu calendars.`	`504 Year 504 (DIV) was a leap year starting on Thursday (link will display the full calendar) of the Julian calendar. At the time, it was known as the Year of the Consulship of Nicomachus without colleague (or, less frequently, year 1257 "Ab urbe condita"). The denomination 504 for this year has been used since the early medieval period, when the Anno Domini calendar era became the prevalent method in Europe for naming years.`
`When is the dialectical method used?`	Dialectic Dialectic or dialectics (Greek: διαλεκτική, dialektikḗ; related to dialogue), also known as the dialectical method, is at base a discourse between two or more people holding different points of view about a subject but wishing to establish the truth through reasoned arguments. Dialectic resembles debate, but the concept excludes subjective elements such as emotional appeal and the modern pejorative sense of rhetoric.[1][2] Dialectic may be contrasted with the didactic method, wherein one side of the conversation teaches the other. Dialectic is alternatively known as minor logic, as opposed to major logic or critique.	Derek Bentley case Another factor in the posthumous defence was that a "confession" recorded by Bentley, which was claimed by the prosecution to be a "verbatim record of dictated monologue", was shown by forensic linguistics methods to have been largely edited by policemen. Linguist Malcolm Coulthard showed that certain patterns, such as the frequency of the word "then" and the grammatical use of "then" after the grammatical subject ("I then" rather than "then I"), were not consistent with Bentley's use of language (his idiolect), as evidenced in court testimony. These patterns fit better the recorded testimony of the policemen involved. This is one of the earliest uses of forensic linguistics on record.
`What do Grasshoppers eat?`	Grasshopper Grasshoppers are plant-eaters, with a few species at times becoming serious pests of cereals, vegetables and pasture, especially when they swarm in their millions as locusts and destroy crops over wide areas. They protect themselves from predators by camouflage; when detected, many species attempt to startle the predator with a brilliantly-coloured wing-flash while jumping and (if adult) launching themselves into the air, usually flying for only a short distance. Other species such as the rainbow grasshopper have warning coloration which deters predators. Grasshoppers are affected by parasites and various diseases, and many predatory creatures feed on both nymphs and adults. The eggs are the subject of attack by parasitoids and predators.	Groundhog Very often the dens of groundhogs provide homes for other animals including skunks, red foxes, and cottontail rabbits. The fox and skunk feed upon field mice, grasshoppers, beetles and other creatures that destroy farm crops. In aiding these animals, the groundhog indirectly helps the farmer. In addition to providing homes for itself and other animals, the groundhog aids in soil improvement by bringing subsoil to the surface. The groundhog is also a valuable game animal and is considered a difficult sport when hunted in a fair manner. In some parts of Appalachia, they are eaten.

Loss: MultipleNegativesRankingLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "cos_sim"
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 12
per_device_eval_batch_size: 12
gradient_accumulation_steps: 4
num_train_epochs: 1
warmup_ratio: 0.3
bf16: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 12
per_device_eval_batch_size: 12
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 4
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: 1
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.3
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: 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: False
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: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Click to expand

Epoch	Step	Training Loss	sts-dev-896_spearman_cosine	sts-dev-768_spearman_cosine
0.0004	10	2.2049	-	-
0.0009	20	2.3168	-	-
0.0013	30	2.3544	-	-
0.0018	40	2.2519	-	-
0.0022	50	2.1809	-	-
0.0027	60	2.1572	-	-
0.0031	70	2.1855	-	-
0.0036	80	2.5887	-	-
0.0040	90	2.883	-	-
0.0045	100	2.8557	-	-
0.0049	110	2.9356	-	-
0.0053	120	2.8833	-	-
0.0058	130	2.8394	-	-
0.0062	140	2.923	-	-
0.0067	150	2.8191	-	-
0.0071	160	2.8658	-	-
0.0076	170	2.8252	-	-
0.0080	180	2.8312	-	-
0.0085	190	2.7761	-	-
0.0089	200	2.7193	-	-
0.0094	210	2.724	-	-
0.0098	220	2.7484	-	-
0.0102	230	2.7262	-	-
0.0107	240	2.6964	-	-
0.0111	250	2.6676	-	-
0.0116	260	2.6715	-	-
0.0120	270	2.6145	-	-
0.0125	280	2.6191	-	-
0.0129	290	1.9812	-	-
0.0134	300	1.6413	-	-
0.0138	310	1.6126	-	-
0.0143	320	1.3599	-	-
0.0147	330	1.2996	-	-
0.0151	340	1.2654	-	-
0.0156	350	1.9409	-	-
0.0160	360	2.1287	-	-
0.0165	370	1.8442	-	-
0.0169	380	1.6837	-	-
0.0174	390	1.5489	-	-
0.0178	400	1.4382	-	-
0.0183	410	1.4848	-	-
0.0187	420	1.3481	-	-
0.0192	430	1.3467	-	-
0.0196	440	1.3977	-	-
0.0201	450	1.26	-	-
0.0205	460	1.2412	-	-
0.0209	470	1.316	-	-
0.0214	480	1.3501	-	-
0.0218	490	1.2246	-	-
0.0223	500	1.2271	-	-
0.0227	510	1.1871	-	-
0.0232	520	1.1685	-	-
0.0236	530	1.1624	-	-
0.0241	540	1.1911	-	-
0.0245	550	1.1978	-	-
0.0250	560	1.1228	-	-
0.0254	570	1.1091	-	-
0.0258	580	1.1433	-	-
0.0263	590	1.0638	-	-
0.0267	600	1.0515	-	-
0.0272	610	1.175	-	-
0.0276	620	1.0943	-	-
0.0281	630	1.1226	-	-
0.0285	640	0.9871	-	-
0.0290	650	1.0171	-	-
0.0294	660	1.0169	-	-
0.0299	670	0.9643	-	-
0.0303	680	0.9563	-	-
0.0307	690	0.9841	-	-
0.0312	700	1.0349	-	-
0.0316	710	0.8958	-	-
0.0321	720	0.9225	-	-
0.0325	730	0.842	-	-
0.0330	740	0.9104	-	-
0.0334	750	0.8927	-	-
0.0339	760	0.8508	-	-
0.0343	770	0.8835	-	-
0.0348	780	0.9531	-	-
0.0352	790	0.926	-	-
0.0356	800	0.8718	-	-
0.0361	810	0.8261	-	-
0.0365	820	0.8169	-	-
0.0370	830	0.8525	-	-
0.0374	840	0.8504	-	-
0.0379	850	0.7625	-	-
0.0383	860	0.8259	-	-
0.0388	870	0.7558	-	-
0.0392	880	0.7898	-	-
0.0397	890	0.7694	-	-
0.0401	900	0.7429	-	-
0.0405	910	0.6666	-	-
0.0410	920	0.7407	-	-
0.0414	930	0.6665	-	-
0.0419	940	0.7597	-	-
0.0423	950	0.7035	-	-
0.0428	960	0.7166	-	-
0.0432	970	0.6889	-	-
0.0437	980	0.7541	-	-
0.0441	990	0.7175	-	-
0.0446	1000	0.7389	0.6420	0.6403
0.0450	1010	0.7142	-	-
0.0454	1020	0.7301	-	-
0.0459	1030	0.7299	-	-
0.0463	1040	0.6759	-	-
0.0468	1050	0.7036	-	-
0.0472	1060	0.6286	-	-
0.0477	1070	0.595	-	-
0.0481	1080	0.6099	-	-
0.0486	1090	0.6377	-	-
0.0490	1100	0.6309	-	-
0.0495	1110	0.6306	-	-
0.0499	1120	0.557	-	-
0.0504	1130	0.5898	-	-
0.0508	1140	0.5896	-	-
0.0512	1150	0.6399	-	-
0.0517	1160	0.5923	-	-
0.0521	1170	0.5787	-	-
0.0526	1180	0.591	-	-
0.0530	1190	0.5714	-	-
0.0535	1200	0.6047	-	-
0.0539	1210	0.5904	-	-
0.0544	1220	0.543	-	-
0.0548	1230	0.6033	-	-
0.0553	1240	0.5445	-	-
0.0557	1250	0.5217	-	-
0.0561	1260	0.5835	-	-
0.0566	1270	0.5353	-	-
0.0570	1280	0.5887	-	-
0.0575	1290	0.5967	-	-
0.0579	1300	0.5036	-	-
0.0584	1310	0.5915	-	-
0.0588	1320	0.5719	-	-
0.0593	1330	0.5238	-	-
0.0597	1340	0.5647	-	-
0.0602	1350	0.538	-	-
0.0606	1360	0.5457	-	-
0.0610	1370	0.5169	-	-
0.0615	1380	0.4967	-	-
0.0619	1390	0.4864	-	-
0.0624	1400	0.5133	-	-
0.0628	1410	0.5587	-	-
0.0633	1420	0.4691	-	-
0.0637	1430	0.5186	-	-
0.0642	1440	0.4907	-	-
0.0646	1450	0.5281	-	-
0.0651	1460	0.4741	-	-
0.0655	1470	0.4452	-	-
0.0659	1480	0.4771	-	-
0.0664	1490	0.4289	-	-
0.0668	1500	0.4551	-	-
0.0673	1510	0.4558	-	-
0.0677	1520	0.5159	-	-
0.0682	1530	0.4296	-	-
0.0686	1540	0.4548	-	-
0.0691	1550	0.4439	-	-
0.0695	1560	0.4295	-	-
0.0700	1570	0.4466	-	-
0.0704	1580	0.4717	-	-
0.0708	1590	0.492	-	-
0.0713	1600	0.4566	-	-
0.0717	1610	0.4451	-	-
0.0722	1620	0.4715	-	-
0.0726	1630	0.4573	-	-
0.0731	1640	0.3972	-	-
0.0735	1650	0.5212	-	-
0.0740	1660	0.4381	-	-
0.0744	1670	0.4552	-	-
0.0749	1680	0.4767	-	-
0.0753	1690	0.4398	-	-
0.0757	1700	0.4801	-	-
0.0762	1710	0.3751	-	-
0.0766	1720	0.4407	-	-
0.0771	1730	0.4305	-	-
0.0775	1740	0.3938	-	-
0.0780	1750	0.4748	-	-
0.0784	1760	0.428	-	-
0.0789	1770	0.404	-	-
0.0793	1780	0.4261	-	-
0.0798	1790	0.359	-	-
0.0802	1800	0.4422	-	-
0.0807	1810	0.4748	-	-
0.0811	1820	0.4352	-	-
0.0815	1830	0.4032	-	-
0.0820	1840	0.4124	-	-
0.0824	1850	0.4486	-	-
0.0829	1860	0.429	-	-
0.0833	1870	0.4189	-	-
0.0838	1880	0.3658	-	-
0.0842	1890	0.4297	-	-
0.0847	1900	0.4215	-	-
0.0851	1910	0.3726	-	-
0.0856	1920	0.3736	-	-
0.0860	1930	0.4287	-	-
0.0864	1940	0.4402	-	-
0.0869	1950	0.4353	-	-
0.0873	1960	0.3622	-	-
0.0878	1970	0.3557	-	-
0.0882	1980	0.4107	-	-
0.0887	1990	0.3982	-	-
0.0891	2000	0.453	0.7292	0.7261
0.0896	2010	0.3971	-	-
0.0900	2020	0.4374	-	-
0.0905	2030	0.4322	-	-
0.0909	2040	0.3945	-	-
0.0913	2050	0.356	-	-
0.0918	2060	0.4182	-	-
0.0922	2070	0.3694	-	-
0.0927	2080	0.3989	-	-
0.0931	2090	0.4237	-	-
0.0936	2100	0.3961	-	-
0.0940	2110	0.4264	-	-
0.0945	2120	0.3609	-	-
0.0949	2130	0.4154	-	-
0.0954	2140	0.3661	-	-
0.0958	2150	0.3328	-	-
0.0962	2160	0.3456	-	-
0.0967	2170	0.3478	-	-
0.0971	2180	0.3339	-	-
0.0976	2190	0.3833	-	-
0.0980	2200	0.3238	-	-
0.0985	2210	0.3871	-	-
0.0989	2220	0.4009	-	-
0.0994	2230	0.4115	-	-
0.0998	2240	0.4024	-	-
0.1003	2250	0.35	-	-
0.1007	2260	0.3649	-	-
0.1011	2270	0.3615	-	-
0.1016	2280	0.3898	-	-
0.1020	2290	0.3866	-	-
0.1025	2300	0.3904	-	-
0.1029	2310	0.3321	-	-
0.1034	2320	0.3803	-	-
0.1038	2330	0.3831	-	-
0.1043	2340	0.403	-	-
0.1047	2350	0.3803	-	-
0.1052	2360	0.3463	-	-
0.1056	2370	0.3987	-	-
0.1060	2380	0.3731	-	-
0.1065	2390	0.353	-	-
0.1069	2400	0.3166	-	-
0.1074	2410	0.3895	-	-
0.1078	2420	0.4025	-	-
0.1083	2430	0.3798	-	-
0.1087	2440	0.2991	-	-
0.1092	2450	0.3094	-	-
0.1096	2460	0.3669	-	-
0.1101	2470	0.3412	-	-
0.1105	2480	0.3697	-	-
0.1110	2490	0.369	-	-
0.1114	2500	0.3393	-	-
0.1118	2510	0.4232	-	-
0.1123	2520	0.3445	-	-
0.1127	2530	0.4165	-	-
0.1132	2540	0.3721	-	-
0.1136	2550	0.3476	-	-
0.1141	2560	0.2847	-	-
0.1145	2570	0.3609	-	-
0.1150	2580	0.3017	-	-
0.1154	2590	0.374	-	-
0.1159	2600	0.3365	-	-
0.1163	2610	0.393	-	-
0.1167	2620	0.3623	-	-
0.1172	2630	0.3538	-	-
0.1176	2640	0.3206	-	-
0.1181	2650	0.3962	-	-
0.1185	2660	0.3087	-	-
0.1190	2670	0.3482	-	-
0.1194	2680	0.3616	-	-
0.1199	2690	0.3955	-	-
0.1203	2700	0.3915	-	-
0.1208	2710	0.3782	-	-
0.1212	2720	0.3576	-	-
0.1216	2730	0.3544	-	-
0.1221	2740	0.3572	-	-
0.1225	2750	0.3107	-	-
0.1230	2760	0.3579	-	-
0.1234	2770	0.3571	-	-
0.1239	2780	0.3694	-	-
0.1243	2790	0.3674	-	-
0.1248	2800	0.3373	-	-
0.1252	2810	0.3362	-	-
0.1257	2820	0.3225	-	-
0.1261	2830	0.3609	-	-
0.1265	2840	0.3681	-	-
0.1270	2850	0.4059	-	-
0.1274	2860	0.3047	-	-
0.1279	2870	0.3446	-	-
0.1283	2880	0.3507	-	-
0.1288	2890	0.3124	-	-
0.1292	2900	0.3712	-	-
0.1297	2910	0.3394	-	-
0.1301	2920	0.3869	-	-
0.1306	2930	0.3449	-	-
0.1310	2940	0.3752	-	-
0.1314	2950	0.3341	-	-
0.1319	2960	0.3329	-	-
0.1323	2970	0.36	-	-
0.1328	2980	0.3788	-	-
0.1332	2990	0.3834	-	-
0.1337	3000	0.3426	0.7603	0.7590

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.3.0
Transformers: 4.46.2
PyTorch: 2.1.0+cu118
Accelerate: 1.1.1
Datasets: 3.1.0
Tokenizers: 0.20.3

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",
}

MultipleNegativesRankingLoss

@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}
}