Search public pages, research tools, and SAVRN solutions.
Jev-Omni is an open-weight model for text classification from Wiktor Ryznar, released under Apache License 2.0. Its published files total 71.6 GB.
A multimodal decision classifier for text, images, audio and video. Supply a question and options; receive a probability for each option—not a generated explanation. Built on Gemma 4 12B IT, with a 30,000-question fine-tuning run. Merged-model results.
By Wiktor Ryznar, published under apache-2.0, revision 2733f943232d.
A multimodal decision classifier for text, images, audio and video. Supply a question and options; receive a probability for each option—not a generated explanation. Built on Gemma 4 12B IT, with a 30,000-question fine-tuning run. Merged-model results. ¹Equal-weight scenario/group average. Reference scores are officially reported by their developers and may use different evaluation protocols. Jev-Omni cost uses its recorded input tokens at OpenRouter's Gemma 3 12B input rate ($0.05/M); its classifier generates no output tokens. Both classifiers are priced one call per question: a classifier answers one question at a time, so the state is re-sent for each of them and there is no discount for…
A multimodal decision classifier for text, images, audio and video. Supply a question and options; receive a probability for each option—not a generated explanation.
Built on Gemma 4 12B IT, with a 30,000-question fine-tuning run.
| Benchmark | Accuracy¹ | Micro accuracy |
|---|---|---|
| DecisionBench Medium · 80 scenarios / 293 questions | 87.57% | 86.01% |
| JevBench · matched 195 groups / 231 decisions | 86.15% | 87.45% |
| MMAU · 1,000 questions | — | 63.10% |
| MVBench · 14 evaluated tasks / 2,786 questions | 53.10% | 53.09% |
Merged-model results. ¹Equal-weight scenario/group average.
| Model | Parameters | MMAU | MVBench | Modalities |
|---|---|---|---|---|
| Jev-Omni | 12B | 63.10% | 53.10% | Text, image, audio, video |
| Inkling | 975B total / 41B active | 77.20% | — | Text, image, audio |
| Qwen3.5-397B-A17B | 397B total / 17B active | — | 77.60% | Text, image, video |
Reference scores are officially reported by their developers and may use different evaluation protocols.
Jev-Omni cost uses its recorded input tokens at OpenRouter's Gemma 3 12B input rate ($0.05/M); its classifier generates no output tokens.
Both classifiers are priced one call per question: a classifier answers one question at a time, so the state is re-sent for each of them and there is no discount for asking several at once. The three chat models are priced one call per state, with all of that state's questions together, which is their own cheapest shape. Splitting a state into per-question calls multiplies input tokens by 2.82x on this set, so pricing the chat models the same way would move all three marks right by roughly that factor without changing the order.
CUDA GPU required. FP32 weights use about 50 GB before runtime overhead; inference uses BF16 autocast.
pip install -r https://huggingface.co/akhilaaa3/Jev-Omni/resolve/main/requirements.txt
# ffmpeg is also required for audio input
File: requirements.txt
from huggingface_hub import snapshot_download
path = snapshot_download("akhilaaa3/Jev-Omni")
import sys; sys.path.insert(0, path)
from jev_omni import load_jev_omni
classifier = load_jev_omni()
result = classifier.predict(
state="The meeting starts at 10 AM. It is now 9 AM.",
question="Has the meeting started?",
options=["Yes", "No"],
)
print(result)
For an image, audio file or video, add media="/path/to/file" and modality="image", "audio" or "video". The loader downloads the original Gemma 4 multimodal components automatically. Audio is capped at 30 seconds; video uses 16 frames.
Warm H200 inference: 83 ms (~2k-token text), 26 ms (image), 31 ms (13-second audio), 504 ms (16-frame video). Medians over 20 optimized-backend requests; preprocessing and network time are extra.
Best supported at ≤20 options. The head accepts 256, but quality above 20 is not established.
Apache-2.0, following Gemma 4. Dataset rights remain separate.
Medium ECE: 0.0400 (10 bins; lower is better). The graph uses five bins for readability.
Note: Jev-Omni is an independent open model that implements the typed-decision interface — noul (yes/no), choice and score questions answered with calibrated probabilities. It is not affiliated with, endorsed by, sponsored by, or derived from TypeSafe AI or its Jev model, and nothing in it was trained on Jev output.
42 files, 71.6 GB in total. The weights are 16 files totalling 71.6 GB in pt, safetensors.
| File | Type | Size | SHA-256 |
|---|---|---|---|
| backbone/model-00001-of-00013.safetensors | Weights | 4.0 GB | 9a8f104173b8 |
| backbone/model-00002-of-00013.safetensors | Weights | 3.8 GB | 385935464075 |
| backbone/model-00003-of-00013.safetensors | Weights | 3.9 GB | ea3bece64e07 |
| backbone/model-00004-of-00013.safetensors | Weights | 3.9 GB | 3a9f78f87ae0 |
| backbone/model-00005-of-00013.safetensors | Weights | 4.0 GB | 02ba5ca16c7e |
| backbone/model-00006-of-00013.safetensors | Weights | 3.9 GB | be8871ea8177 |
| backbone/model-00007-of-00013.safetensors | Weights | 3.9 GB | b4b4a1ad4b76 |
| backbone/model-00008-of-00013.safetensors | Weights | 4.0 GB | a9db6226909d |
| backbone/model-00009-of-00013.safetensors | Weights | 3.9 GB | f4acf5e7e4d9 |
| backbone/model-00010-of-00013.safetensors | Weights | 3.9 GB | a1e82c624be3 |
| backbone/model-00011-of-00013.safetensors | Weights | 4.0 GB | 518067119b85 |
| backbone/model-00012-of-00013.safetensors | Weights | 4.0 GB | 333ddb85529c |
| backbone/model-00013-of-00013.safetensors | Weights | 471.9 MB | 2eb18abde4f2 |
| head.pt | Weights | 4.0 MB | 8c81edecd733 |
| runtime_buffers.pt | Weights | 19.7 KB | 7310daabb66c |
| unified/model.safetensors | Weights | 23.9 GB | d78782f3b9c3 |
| backbone/config.json | Configuration | 3.1 KB | — |
| backbone/model.safetensors.index.json | Configuration | 51.4 KB | — |
| decision_config.json | Configuration | 1.2 KB | — |
| example.py | Configuration | 389 B | — |
| jev_omni.py | Configuration | 6.2 KB | — |
| load_model.py | Configuration | 3.4 KB | — |
| sha256.json | Configuration | 3.2 KB | — |
| unified/config.json | Configuration | 4.0 KB | — |
| unified/generation_config.json | Configuration | 260 B | — |
| unified/processor_config.json | Configuration | 1.4 KB | — |
| unified/verification_unified.json | Configuration | 1.3 KB | — |
| verification.json | Configuration | 350 B | — |
| README.md | Documentation | 4.1 KB | — |
| assets/medium-accuracy.png | Other | 79.9 KB | 9b7de696d8b8 |
| assets/medium-accuracy.svg | Other | 4.0 KB | — |
| assets/medium-calibration.png | Other | 140.4 KB | fb1b20ef9250 |
| assets/medium-calibration.svg | Other | 5.0 KB | — |
| assets/the-wall-23s.mp4 | Other | 20.2 MB | 3f1f733188c3 |
| chat_template.jinja | Other | 18.7 KB | — |
| requirements.txt | Other | 118 B | — |
| unified/chat_template.jinja | Other | 18.7 KB | — |
| .gitattributes | Repository | 1.8 KB | — |
| tokenizer.json | Tokenizer | 32.2 MB | cc8d3a0ce364 |
| tokenizer_config.json | Tokenizer | 3.7 KB | — |
| unified/tokenizer.json | Tokenizer | 32.2 MB | cc8d3a0ce364 |
| unified/tokenizer_config.json | Tokenizer | 3.7 KB | — |
Released by Wiktor Ryznar through its official repository on Hugging Face. Read the license.
| Precision | Weights in memory |
|---|---|
| As published | 71.6 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Yes. Jev-Omni 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.
This is a RoBERTa-base model trained on ~124M tweets from January 2018 to December 2021, and finetuned for sentiment analysis with the TweetEval benchmark. The original Twitter-based RoBERTa model can be found here and the original reference paper is TweetEval. This model is suitable for English. 0 -> Negative; 1 -> Neutral; 2 -> Positive This sentiment analysis model has been integrated into TweetNLP. You can access the demo here.
FinBERT is a BERT model pre-trained on financial communication text. The purpose is to enhance financial NLP research and practice. It is trained on the following three financial communication corpus. The total corpora size is 4.9B tokens. More technical details on FinBERT: Click Link This released finbert-tone model is the FinBERT model fine-tuned on 10,000 manually annotated (positive, negative, neutral) sentences from analyst reports. This model achieves superior performance on financial tone analysis task. If you are simply interested in using FinBERT for financial tone analysis, give it a try. If you use the model in your academic work, please cite the following paper: Huang, Allen H.…
https://huggingface.co/cross-encoder/ms-marco-MiniLM-L-6-v2 with ONNX weights to be compatible with Transformers.js. If you haven't already, you can install the Transformers.js JavaScript library from NPM using: Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using Optimum and structuring your repo like this one (with ONNX weights located in a subfolder named onnx).
This is a multilingual XLM-roBERTa-base model trained on ~198M tweets and finetuned for sentiment analysis. The sentiment fine-tuning was done on 8 languages (Ar, En, Fr, De, Hi, It, Sp, Pt) but it can be used for more languages (see paper for details). This model has been integrated into the TweetNLP library.
The huggingface models currently give different results to the detoxify library (see issue here). For the most up to date models we recommend using the models from https://github.com/unitaryai/detoxify Trained models & code to predict toxic comments on 3 Jigsaw challenges: Toxic comment classification, Unintended Bias in Toxic comments, Multilingual toxic comment classification. Built by Laura Hanu at Unitary, where we are working to stop harmful content online by interpreting visual content in context. Score not directly comparable since it is obtained on the validation set provided and not on the test set. To update when the test labels are made available. It is also noteworthy to mention…
With this model, you can classify emotions in English text data. The model was trained on 6 diverse datasets (see Appendix below) and predicts Ekman's 6 basic emotions, plus a neutral class: 1) anger 2) disgust 3) fear 4) joy 5) neutral 6) sadness 7) surprise The model is a fine-tuned checkpoint of DistilRoBERTa-base. For a 'non-distilled' emotion model, please refer to the model card of the RoBERTa-large version. a) Run emotion model with 3 lines of code on single text example using Hugging Face's pipeline command on Google Colab: b) Run emotion model on multiple examples and full datasets (e.g.,.csv files) on Google Colab: Please reach out to [email protected] if you have any…