Search public pages, research tools, and SAVRN solutions.
loupe-1.0 is an open-weight model for text classification from Cong Tuan Tong, released under Apache License 2.0. Its published files total 731.4 MB.
A LoRA adapter for Qwen/Qwen3.5-4B that makes typed decisions about a document in one forward pass and returns calibrated probabilities over the options: a System-1 decision engine, fast judgment rather than generated reasoning.
By Cong Tuan Tong, published under apache-2.0, revision 216e5eaa4af2.
A LoRA adapter for Qwen/Qwen3.5-4B that makes typed decisions about a document in one forward pass and returns calibrated probabilities over the options: a System-1 decision engine, fast judgment rather than generated reasoning. Served with the code at github.com/tuantong/loupe-serving, which reads the answer from the next-token probabilities of the option labels; nothing is generated. An exact average (model soup) of LoRA adapters on the attention, linear-attention and MLP projections of all 32 layers (three seeds of one recipe, a continuation of that recipe on written rationales, and a run whose targets include a second teacher), stored as one rank-96 adapter. Merge it into the base…
A LoRA adapter for Qwen/Qwen3.5-4B that makes typed decisions about a document in one forward pass and returns calibrated probabilities over the options: a System-1 decision engine, fast judgment rather than generated reasoning. Served with the code at github.com/tuantong/loupe-serving, which reads the answer from the next-token probabilities of the option labels; nothing is generated.
An exact average (model soup) of LoRA adapters on the attention, linear-attention and MLP projections of
all 32 layers (three seeds of one recipe, a continuation of that recipe on written rationales, and a run whose
targets include a second teacher), stored as one rank-96 adapter. Merge it into the base weights for
serving. It was trained with the compact prompt template, so serve it with LOUPE_PROMPT_STYLE=compact.
calibration.json reads answers to prompts of 600 tokens or more at temperature 0.7 and shorter ones at
1; the temperature was fitted on our own held-out cases, not on benchmark items.
Public tiers of a System-1 decision benchmark (typed decisions over documents, many with a tempting wrong reading):
| Tier | Items | Accuracy |
|---|---|---|
| Easy | 48 | 100.0% |
| Standard | 72 | 100.0% |
| Hard | 111 | 78.4% |
Hard-tier ECE 0.060; mean total variation distance to the exact distributions of the probability items 0.162. Serial loopback latency on an RTX PRO 6000 Blackwell with the CUDA-graph backend and the adapter merged: p50 13.9 ms, p95 15.8 ms.
English only. It decides from what the document says and does not look anything up. Temporal arithmetic over long documents, ranked-rule tradeoffs and long policies remain its weakest families.
Apache-2.0, like the base model (Qwen/Qwen3.5-4B) and the serving code. The training data was written by DeepSeek V4.1 Flash, with targets from Qwen3.8-27B (Apache-2.0) and DeepSeek V4.1 Flash.
5 files, 731.4 MB in total. The weights are 1 file totalling 731.4 MB in safetensors.
| File | Type | Size | SHA-256 |
|---|---|---|---|
| adapter_model.safetensors | Weights | 731.4 MB | 585ff58da0dc |
| adapter_config.json | Configuration | 1.6 KB | — |
| calibration.json | Configuration | 156 B | — |
| README.md | Documentation | 3.1 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
Released by Cong Tuan Tong through its official repository on Hugging Face. Read the license.
| Precision | Weights in memory |
|---|---|
| As published | 731.4 MB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Yes. loupe-1.0 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…