Rankle-Matching/setfit-importance-v0.2.0

SetFit with sentence-transformers/paraphrase-mpnet-base-v2

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/paraphrase-mpnet-base-v2 as the Sentence Transformer embedding model. A LogisticRegression instance is used for classification.

The model has been trained using an efficient few-shot learning technique that involves:

1. Fine-tuning a Sentence Transformer with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.

Model Details

Model Description

• Model Type: SetFit
• Sentence Transformer body: sentence-transformers/paraphrase-mpnet-base-v2
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 512 tokens
• Number of Classes: 2 classes

Model Sources

• Repository: SetFit on GitHub
• Paper: Efficient Few-Shot Learning Without Prompts
• Blogpost: SetFit: Efficient Few-Shot Learning Without Prompts

Model Labels

Label	Examples
required	'We Are HiringSoftware Engineer (.NET)9 Sri Lanka19uteverdentra1+ years of experience in .NET/C# development with solid SQL knowledgeFamiliar with REST APIs, cloud platforms (Azure/AWS), and DevOps toolsStrong problem-solving skills and a colla' 'nts:. Preferably Malaysian, Vietnamese, or Thai nationals currently living inSri Lanka or still residing in their home country· Basic knowledge of HR operations and admin tasks· Good communication and interpersonal skills. Computer literacy (email, spreadsheets, documentation). Ability to maintain confident' 'Malaysian, Vietnamese, or Thai nationals currently living inSri Lanka or still residing in their home country· Basic knowledge of HR operations and admin tasks· Good communication and interpersonal skills. Computer literacy (email, spreadsheets, documentation). Ability to maintain confidentiality and work'
nice-to-have	'me country· Basic knowledge of HR operations and admin tasks· Good communication and interpersonal skills. Computer literacy (email, spreadsheets, documentation). Ability to maintain confidentiality and work under minimal supervision· Fluent in English (spoken & written)What We Offer:Salary: Paid in USDV' "riented and able to manage multiple tasks effectively.- Familiarity with business analysis techniques or tools is a plus.- Familiarity working with AI tools (Copilot, ChatGPT, etc.) is a plus.- Self-starting and independent with minimum supervision.If you're excited about this opportunity, we'd lo" 'd, PowerPoint).- Strong written and verbal communication skills.- Detail-oriented and able to manage multiple tasks effectively.- Familiarity with business analysis techniques or tools is a plus.- Familiarity working with AI tools (Copilot, ChatGPT, etc.) is a plus.- Self-starting and independent with minimum s'

Uses

Direct Use for Inference

First install the SetFit library:

pip install setfit

Then you can load this model and run inference.

from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("setfit_model_id")
# Run inference
preds = model("Computer experience including: Lawson; Excel; Word; PowerPoint; Allscripts PM; Allscripts HR.")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	7	29.4593	160

Label	Training Sample Count
nice-to-have	53
required	82

Training Hyperparameters

• batch_size: (8, 8)
• num_epochs: (4, 4)
• max_steps: -1
• sampling_strategy: oversampling
• body_learning_rate: (2e-05, 1e-05)
• head_learning_rate: 0.01
• loss: CosineSimilarityLoss
• distance_metric: cosine_distance
• margin: 0.25
• end_to_end: False
• use_amp: False
• warmup_proportion: 0.1
• l2_weight: 0.01
• seed: 42
• eval_max_steps: -1
• load_best_model_at_end: True

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0008	1	0.3838	-
0.0414	50	0.2635	-
0.0827	100	0.244	-
0.1241	150	0.2378	-
0.1654	200	0.2173	-
0.2068	250	0.1599	-
0.2481	300	0.052	-
0.2895	350	0.0109	-
0.3309	400	0.0016	-
0.3722	450	0.0008	-
0.4136	500	0.0004	-
0.4549	550	0.0002	-
0.4963	600	0.0002	-
0.5376	650	0.0001	-
0.5790	700	0.0001	-
0.6203	750	0.0001	-
0.6617	800	0.0001	-
0.7031	850	0.0001	-
0.7444	900	0.0001	-
0.7858	950	0.0006	-
0.8271	1000	0.0084	-
0.8685	1050	0.0002	-
0.9098	1100	0.0001	-
0.9512	1150	0.0001	-
0.9926	1200	0.0001	-
1.0	1209	-	0.2705
1.0339	1250	0.0001	-
1.0753	1300	0.0001	-
1.1166	1350	0.0	-
1.1580	1400	0.0001	-
1.1993	1450	0.002	-
1.2407	1500	0.0005	-
1.2821	1550	0.0001	-
1.3234	1600	0.0	-
1.3648	1650	0.0	-
1.4061	1700	0.0	-
1.4475	1750	0.0	-
1.4888	1800	0.0	-
1.5302	1850	0.0	-
1.5715	1900	0.0	-
1.6129	1950	0.0	-
1.6543	2000	0.0466	-
1.6956	2050	0.016	-
1.7370	2100	0.0041	-
1.7783	2150	0.0001	-
1.8197	2200	0.0001	-
1.8610	2250	0.0	-
1.9024	2300	0.0001	-
1.9438	2350	0.0012	-
1.9851	2400	0.0	-
2.0	2418	-	0.3016
2.0265	2450	0.0	-
2.0678	2500	0.0	-
2.1092	2550	0.0	-
2.1505	2600	0.0	-
2.1919	2650	0.0	-
2.2333	2700	0.0	-
2.2746	2750	0.0	-
2.3160	2800	0.0	-
2.3573	2850	0.0	-
2.3987	2900	0.0	-
2.4400	2950	0.0	-
2.4814	3000	0.0	-
2.5227	3050	0.0	-
2.5641	3100	0.0	-
2.6055	3150	0.0	-
2.6468	3200	0.0	-
2.6882	3250	0.0	-
2.7295	3300	0.0	-
2.7709	3350	0.0	-
2.8122	3400	0.0	-
2.8536	3450	0.0	-
2.8950	3500	0.0	-
2.9363	3550	0.0	-
2.9777	3600	0.0	-
3.0	3627	-	0.3102
3.0190	3650	0.0	-
3.0604	3700	0.0	-
3.1017	3750	0.0	-
3.1431	3800	0.0	-
3.1844	3850	0.0	-
3.2258	3900	0.0	-
3.2672	3950	0.0	-
3.3085	4000	0.0	-
3.3499	4050	0.0	-
3.3912	4100	0.0	-
3.4326	4150	0.0	-
3.4739	4200	0.0	-
3.5153	4250	0.0	-
3.5567	4300	0.0	-
3.5980	4350	0.0	-
3.6394	4400	0.0	-
3.6807	4450	0.0	-
3.7221	4500	0.0	-
3.7634	4550	0.0	-
3.8048	4600	0.0	-
3.8462	4650	0.0	-
3.8875	4700	0.0	-
3.9289	4750	0.0	-
3.9702	4800	0.0	-
4.0	4836	-	0.3119

Framework Versions

• Python: 3.12.3
• SetFit: 1.1.3
• Sentence Transformers: 5.2.0
• Transformers: 4.57.3
• PyTorch: 2.5.1+cu124
• Datasets: 4.4.2
• Tokenizers: 0.22.2

Citation

BibTeX

@article{https://doi.org/10.48550/arxiv.2209.11055,
    doi = {10.48550/ARXIV.2209.11055},
    url = {https://arxiv.org/abs/2209.11055},
    author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
    keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
    title = {Efficient Few-Shot Learning Without Prompts},
    publisher = {arXiv},
    year = {2022},
    copyright = {Creative Commons Attribution 4.0 International}
}