nli-deberta-v3-small: Open-Weight Model

Runs On

What it takes to serve nli-deberta-v3-small (142M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

Precision	Weights	Memory needed	Cheapest setup	Per hour	Also fits
16-bit	0.3 GB	0.3 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
8-bit	0.1 GB	0.2 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
4-bit	0.1 GB	0.1 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

SAVRN's Notes on nli-deberta-v3-small

For every sentence pair you hand it, this cross-encoder returns three scores, one each for contradiction, entailment and neutral, which is the basis for its zero-shot classification use. Six layers and 142M parameters keep memory small: 0.3 GB at 16-bit, 0.2 GB at 8-bit, 0.1 GB at 4-bit. The cheapest setup on our Index is one MI300X with 192 GB at $1.85 per hour on-demand, which leaves almost the whole card idle, so share it. The download is another matter: 20 files and 4.45 GB across safetensors, ONNX and PyTorch formats.

Apache 2.0 terms apply: commercial use, modification and redistribution, with the notices and NOTICE file kept, significant changes stated, and an express patent grant included. Check the 512 token context and the lineage: it derives from microsoft/deberta-v3-small and was trained on SNLI and MultiNLI. No evaluation figures are listed here, so test it on your own pairs.

SAVRN Research, 2026-09-18

Model Card

By Sentence Transformers - Cross-Encoders, published under apache-2.0, revision fa2804872c3b.

Cross-Encoder for Natural Language Inference

This model was trained using SentenceTransformers Cross-Encoder class. This model is based on microsoft/deberta-v3-small

Training Data

The model was trained on the SNLI and MultiNLI datasets. For a given sentence pair, it will output three scores corresponding to the labels: contradiction, entailment, neutral.

Performance

Accuracy on SNLI-test dataset: 91.65
Accuracy on MNLI mismatched set: 87.55

For futher evaluation results, see SBERT.net - Pretrained Cross-Encoder.

Usage

Pre-trained models can be used like this:

from sentence_transformers import CrossEncoder
model = CrossEncoder('cross-encoder/nli-deberta-v3-small')
scores = model.predict([('A man is eating pizza', 'A man eats something'), ('A black race car starts up in front of a crowd of people.', 'A man is driving down a lonely road.')])

#Convert scores to labels
label_mapping = ['contradiction', 'entailment', 'neutral']
labels = [label_mapping[score_max] for score_max in scores.argmax(axis=1)]

Usage with Transformers AutoModel

You can use the model also directly with Transformers library (without SentenceTransformers library):

Read the full model card (254 words)

Configuration

Architecture: DebertaV2ForSequenceClassification
Context length (tokens): 512
Layers: 6
Hidden size: 768
Feed-forward size: 3,072
Attention heads: 12
Vocabulary size: 128,100
Stored precision: float32
Model type: deberta-v2

Identity and Version

Repository: cross-encoder/nli-deberta-v3-small
Publisher: Sentence Transformers - Cross-Encoders
Task: Zero-shot classification
Modality: Text
Library: sentence-transformers
Parameters: 142M parameters
Languages: en
Revision: fa2804872c3b4bd748f38c0185cc85775361e735
First published: 2022-03-02
Last updated: 2025-04-15

Files and Weights

20 files, 4.5 GB in total. The weights are 11 files totalling 4.4 GB in bin, onnx, safetensors.

Weights11 files · 4.4 GB

Configuration3 files · 1.4 KB

Tokenizer2 files · 8.7 MB

Documentation1 file · 2.9 KB

Other2 files · 2.5 MB

Repository1 file · 1.2 KB

Every file

File	Type	Size	SHA-256
model.safetensors	Weights	567.6 MB	ebc79588dd73
onnx/model.onnx	Weights	568.0 MB	59fd8dd78926
onnx/model_O1.onnx	Weights	585.2 MB	2755d9ae55f8
onnx/model_O2.onnx	Weights	585.1 MB	baabab86ef26
onnx/model_O3.onnx	Weights	585.1 MB	5791ade52d31
onnx/model_O4.onnx	Weights	292.7 MB	0f7bd2134d37
onnx/model_qint8_arm64.onnx	Weights	172.5 MB	a1da6815c8da
onnx/model_qint8_avx512.onnx	Weights	172.5 MB	a1da6815c8da
onnx/model_qint8_avx512_vnni.onnx	Weights	172.5 MB	a1da6815c8da
onnx/model_quint8_avx2.onnx	Weights	172.5 MB	03c2221313dc
pytorch_model.bin	Weights	567.6 MB	2930841dcaa8
added_tokens.json	Configuration	26 B	—
config.json	Configuration	1.1 KB	—
special_tokens_map.json	Configuration	301 B	—
README.md	Documentation	2.9 KB	—
CESoftmaxAccuracyEvaluator_AllNLI-dev_results.csv	Other	678 B	—
spm.model	Other	2.5 MB	c679fbf93643
.gitattributes	Repository	1.2 KB	—
tokenizer.json	Tokenizer	8.7 MB	—
tokenizer_config.json	Tokenizer	1.3 KB	—

License and Download

License: apache-2.0
Access: Open weights, no gate
Download size: 4.4 GB

Download from Sentence Transformers - Cross-Encoders

Released by Sentence Transformers - Cross-Encoders through its official repository on Hugging Face. Read the license.

Built From

Derived from microsoft/deberta-v3-small
Quantized from microsoft/deberta-v3-small
Trained on (disclosed) nyu-mll/multi_nli
Trained on (disclosed) stanfordnlp/snli

Memory Requirements

Precision	Weights in memory
As published	4.4 GB
16-bit	0.3 GB
8-bit	0.1 GB
4-bit	0.1 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Built on This Model

Quantized fromnli-deberta-v3-small
Derived fromnli-deberta-v3-small

Questions About nli-deberta-v3-small

How much GPU memory does nli-deberta-v3-small need?

About 0.3 GB at 16-bit and 0.1 GB at 4-bit: the weights (142M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run nli-deberta-v3-small on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use nli-deberta-v3-small commercially?

Yes. nli-deberta-v3-small is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

What is nli-deberta-v3-small's context length?

512 tokens, from the maximum position embeddings in its published configuration.

Similar Models

Model · Zero-shot classification

deberta-small-long-nli

Tasksource

DeBERTa-v3-small with context length of 1680 tokens fine-tuned on tasksource for 250k steps. I oversampled long NLI tasks (ConTRoL, doc-nli). Training data include HelpSteer v1/v2, logical reasoning tasks (FOLIO, FOL-nli, LogicNLI...), OASST, hh/rlhf, linguistics oriented NLI tasks, tasksource-dpo, fact verification tasks. This model is suitable for long context NLI or as a backbone for reward models or classifiers fine-tuning. This checkpoint has strong zero-shot validation performance on many tasks (e.g. 70% on WNLI), and can be used for: - Zero-shot entailment-based classification for arbitrary labels [ZS]. - Natural language inference [NLI] - Further fine-tuning on a new task or…

Open weights apache-2.0 142M parameters 1,680 tokens transformers

View model

Model · Zero-shot classification

gliclass-small-v1.0

Knowledgator Engineering

This is an efficient zero-shot classifier inspired by GLiNER work. It demonstrates the same performance as a cross-encoder while being more compute-efficient because classification is done at a single forward path. It can be used for topic classification, sentiment analysis and as a reranker in RAG pipelines. The model was trained on synthetic data and can be used in commercial applications. First of all, you need to install GLiClass library: Than you need to initialize a model and a pipeline: Below, you can see the F1 score on several text classification datasets. All tested models were not fine-tuned on those datasets and were tested in a zero-shot setting. Below you can find a comparison…

Open weights apache-2.0 144M parameters transformers

View model

Model · Zero-shot classification

nli-deberta-base

Sentence Transformers - Cross-Encoders

This model was trained using SentenceTransformers Cross-Encoder class. The model was trained on the SNLI and MultiNLI datasets. For a given sentence pair, it will output three scores corresponding to the labels: contradiction, entailment, neutral. For evaluation results, see SBERT.net - Pretrained Cross-Encoder. Pre-trained models can be used like this: You can use the model also directly with Transformers library (without SentenceTransformers library): This model can also be used for zero-shot-classification

Open weights apache-2.0 139M parameters 512 tokens sentence-transformers

View model

Model · Zero-shot classification

ModernBERT-base-nli

Tasksource

ModernBERT multi-task fine-tuned on tasksource NLI tasks, including MNLI, ANLI, SICK, WANLI, doc-nli, LingNLI, FOLIO, FOL-NLI, LogicNLI, Label-NLI and all datasets in the below table). This is the equivalent of an "instruct" version. The model was trained for 200k steps on an Nvidia A30 GPU. It is very good at reasoning tasks (better than llama 3.1 8B Instruct on ANLI and FOLIO), long context reasoning, sentiment analysis and zero-shot classification with new labels. The following table shows model test accuracy. These are the scores for the same single transformer with different classification heads on top. Further gains can be obtained by fine-tuning on a single-task, e.g. SST, but it…

Open weights apache-2.0 150M parameters 2,048 tokens transformers

View model

Model · Zero-shot classification

nli-roberta-base

Sentence Transformers - Cross-Encoders

This model was trained using SentenceTransformers Cross-Encoder class. The model was trained on the SNLI and MultiNLI datasets. For a given sentence pair, it will output three scores corresponding to the labels: contradiction, entailment, neutral. For evaluation results, see SBERT.net - Pretrained Cross-Encoder. Pre-trained models can be used like this: You can use the model also directly with Transformers library (without SentenceTransformers library): This model can also be used for zero-shot-classification

Open weights apache-2.0 125M parameters 514 tokens sentence-transformers

View model

Model · Zero-shot classification

roberta-base-zeroshot-v2.0-c

Moritz Borrett-Laurer (formerly Laurer)

Models in this series are designed for efficient zeroshot classification with the Hugging Face pipeline. These models can do classification without training data and run on both GPUs and CPUs. An overview of the latest zeroshot classifiers is available in my Zeroshot Classifier Collection. The main update of this zeroshot-v2.0 series of models is that several models are trained on fully commercially-friendly data for users with strict license requirements. These models can do one universal classification task: determine whether a hypothesis is "true" or "not true" given a text (entailment vs. notentailment). This task format is based on the Natural Language Inference task (NLI). The task is…

Open weights mit 125M parameters 514 tokens transformers

View model