# phobert-emotion-social by Le Van Huy: Open-Weight Model Source: https://savrn.com/models/phobert-emotion-social Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve phobert-emotion-social (135M 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](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 0.1 GB | 0.2 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 0.1 GB | 0.1 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/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](https://savrn.com/ai-index/pricing/gpus), read Oct 9, 2026. [phobert-emotion-social on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/phobert-emotion-social/gpus) ## Model Card By Le Van Huy, published under mit, revision 0d954dabf3fa. ### PhoBERT Emotion Recognition v1.1 ### Model Description PhoBERT Emotion v1.1 is a fine-tuned sequence classification model based on [vinai/phobert-base-v2](https://huggingface.co/vinai/phobert-base-v2) for 7-class emotion recognition in Vietnamese social media texts: * Enjoyment (Thích thú / Vui vẻ) * Sadness (Buồn bã) * Disgust (Chán ghét / Khinh bỉ) * Anger (Tức giận) * Fear (Sợ hãi) * Surprise (Ngạc nhiên) * Other (Khác / Trung tính) The model is trained on a curated and balanced Vietnamese social corpus (11,193 samples) with anti-overfitting techniques (Cosine Annealing scheduler, early stopping, and weight decay). In addition to standard PyTorch weights, optimized ONNX FP32 and ONNX Dynamic INT8 formats are provided for high-speed, resource-efficient CPU inference in production environments. ### Model Formats & Artifacts | Format | File Path | Model Size | Target Use-Case | | --- | --- | --- | --- | | PyTorch FP32 | model.safetensors | ~515 MB | Standard Hugging Face training & fine-tuning | | ONNX FP32 | onnx/phobert_emotion_fp32.onnx | ~515 MB | 100% precision preservation, ~15% faster CPU inference | | ONNX INT8 | onnx/phobert_emotion_int8.onnx | ~130 MB | Dynamic quantization, 74.8% memory saving, low-latency CPU serving | [Read the full model card (932 words)](https://savrn.com/models/phobert-emotion-social/card) ## Configuration Architecture RobertaForSequenceClassification Context length (tokens) 258 Layers 12 Hidden size 768 Feed-forward size 3,072 Attention heads 12 Vocabulary size 64,001 Model type roberta ## Identity and Version Repository huyleit/phobert-emotion-social Publisher Le Van Huy Task Text classification Modality Text Library transformers Parameters 135M parameters Languages vi Revision 0d954dabf3fa335bbca21def093ceff1ce5503be First published 2026-08-29 Last updated 2026-10-09 ## Files and Weights 14 files, 1.2 GB in total. The weights are 3 files totalling 1.2 GB in onnx, safetensors. Weights3 files · 1.2 GB Configuration3 files · 2.3 KB Tokenizer4 files · 1.8 MB Documentation1 file · 10.3 KB Other2 files · 2.3 MB Repository1 file · 1.5 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model.safetensors | Weights | 540.0 MB | 7a2af3287880 | | onnx/phobert_emotion_fp32.onnx | Weights | 540.3 MB | f2ec4b7b2837 | | onnx/phobert_emotion_int8.onnx | Weights | 135.9 MB | ecd5b2fb8045 | | added_tokens.json | Configuration | 22 B | — | | config.json | Configuration | 1.1 KB | — | | onnx/config.json | Configuration | 1.1 KB | — | | README.md | Documentation | 10.3 KB | — | | bpe.codes | Other | 1.1 MB | — | | onnx/bpe.codes | Other | 1.1 MB | — | | .gitattributes | Repository | 1.5 KB | — | | onnx/tokenizer_config.json | Tokenizer | 1.2 KB | — | | onnx/vocab.txt | Tokenizer | 895.3 KB | — | | tokenizer_config.json | Tokenizer | 1.2 KB | — | | vocab.txt | Tokenizer | 895.3 KB | — | ## License and Download License mit Access Open weights, no gate Download size 1.2 GB [Download from Le Van Huy](https://huggingface.co/huyleit/phobert-emotion-social) Released by Le Van Huy through its official repository on Hugging Face. [Read the license](https://opensource.org/license/mit). ## Built From - Derived from vinai/phobert-base-v2 - Quantized from vinai/phobert-base-v2 - Trained on (disclosed) custom ## Evaluations Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical. | Benchmark | Conditions | Result | Reported by | Revision | Date | | --- | --- | --- | --- | --- | --- | | Vietnamese Social Media Emotion Dataset | Task Text ClassificationMetric AccuracyComparison conditions not established | 0.6432 | [huyleit](https://huggingface.co/huyleit/phobert-emotion-social) Publisher reported | Evaluated revision not stated | — | | Vietnamese Social Media Emotion Dataset | Task Text ClassificationMetric Macro F1Comparison conditions not established | 0.6382 | [huyleit](https://huggingface.co/huyleit/phobert-emotion-social) Publisher reported | Evaluated revision not stated | — | | Vietnamese Social Media Emotion Dataset | Task Text ClassificationMetric Weighted F1Comparison conditions not established | 0.6413 | [huyleit](https://huggingface.co/huyleit/phobert-emotion-social) Publisher reported | Evaluated revision not stated | — | ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 1.2 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. ## Questions About phobert-emotion-social ### How much GPU memory does phobert-emotion-social need? About 0.3 GB at 16-bit and 0.1 GB at 4-bit: the weights (135M parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run phobert-emotion-social 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 phobert-emotion-social commercially? Yes. phobert-emotion-social is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included. ### What is phobert-emotion-social's context length? 258 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Text classification ### [phobert-vi-moderation-v1.1](https://savrn.com/models/phobert-vi-moderation-v1-1) [Le Van Huy](https://savrn.com/model-publishers/huyleit) PhoBERT Moderation v1.1 is a fine-tuned sequence classification model based on vinai/phobert-base-v2 for contextual text moderation in Vietnamese social networks, with a dedicated focus on mental health awareness and crisis intervention. Traditional keyword-based moderation often struggles to distinguish between harmless personal venting and actual malicious hate speech, or fails to detect acute emotional distress. This model implements a nuanced 4-class taxonomy: CLEAN (An toàn / Tích cực) PROFANITYVENTING (Xả bực dọc / Chửi thề vô hại) HATESPEECH (Thù ghét / Công kích xúc phạm) SELFHARMCRISIS (Khủng hoảng tâm lý / Nguy cơ tự hại) Trained on a curated Vietnamese social corpus integrating… Open weights mit 135M parameters 258 tokens transformers [View model](https://savrn.com/models/phobert-vi-moderation-v1-1) Model · Text classification ### [prompt-injection-guard-multilingual](https://savrn.com/models/prompt-injection-guard-multilingual) [Horizon Labs](https://savrn.com/model-publishers/horizon-labs) A compact multilingual detector for prompt injection and jailbreak attempts (large language model guardrails). Given any user prompt, tool output, or document excerpt, it outputs the probability that the text is an attack on the LLM's instructions. Existing public injection detectors are either English-only (e.g. protectai's deberta models) or released under licenses many organizations cannot use (Meta's PromptGuard line). This model is permissively licensed, works in 17 languages, ships with its multilingual training corpus and is evaluated against a public baseline on shared test sets. Recommended threshold: 0.50 — conservative default (EN precision 0.83 on the attack test, safe-prompt… Open weights apache-2.0 135M parameters 512 tokens transformers [View model](https://savrn.com/models/prompt-injection-guard-multilingual) Model · Text classification ### [distilbert-base-multilingual-cased-sentiments-student](https://savrn.com/models/distilbert-base-multilingual-cased-sentiments-student) [Lik Xun Yuan](https://savrn.com/model-publishers/lxyuan) This model is distilled from the zero-shot classification pipeline on the Multilingual Sentiment dataset using this script. In reality the multilingual-sentiment dataset is annotated of course, but we'll pretend and ignore the annotations for the sake of example. Result can be reproduce using the following commands: If you are training this model on Colab, make the following code changes to avoid Out-of-memory error message: - Transformers 4.28.1 - Pytorch 2.0.0+cu118 - Datasets 2.11.0 - Tokenizers 0.13.3 Open weights apache-2.0 135M parameters 512 tokens transformers [View model](https://savrn.com/models/distilbert-base-multilingual-cased-sentiments-student) Model · Text classification ### [turn-detector](https://savrn.com/models/turn-detector) [LiveKit](https://savrn.com/model-publishers/livekit) An open-weights language model for contextually-aware end-of-utterance (EOU) detection in voice AI applications. The model predicts whether a user has finished speaking based on the semantic content of their transcribed speech, providing a critical complement to voice activity detection (VAD) systems. Traditional voice agents rely on voice activity detection (VAD) to determine when a user has finished speaking. VAD works by detecting the presence or absence of speech in an audio signal and applying a silence timer. While effective for detecting pauses, VAD lacks language understanding and frequently causes false positives. For example, a user who says "I need to think about that for a… Open weights other 135M parameters 8,192 tokens transformers [View model](https://savrn.com/models/turn-detector) Model · Text classification ### [prompt-injection-guard-small](https://savrn.com/models/prompt-injection-guard-small) [Horizon Labs](https://savrn.com/model-publishers/horizon-labs) A fast, multilingual classifier that flags prompt injection and jailbreak attempts, both in user messages (direct) and in untrusted content an AI agent reads: emails, web pages, documents, RAG chunks, and tool/API outputs (indirect). their clean counterparts, so it looks for instructions aimed at the AI, not for scary words. - Low false-alarm rate on look-alike benign text: 89.7% on NotInject, 99.5% on OR-Bench-hard. half the size and the same decisions as fp32 on our checks), transformers.js. Labels: SAFE (0) and INJECTION (1). This is the same convention as protectai/deberta-v3-base-prompt-injection-v2, so the model is a drop-in replacement in code and tools built for that one. INJECTION… Open weights apache-2.0 141M parameters 8,192 tokens transformers [View model](https://savrn.com/models/prompt-injection-guard-small) Model · Text classification ### [Julia-1-MLX](https://savrn.com/models/julia-1-mlx) [Zain Merchant](https://savrn.com/model-publishers/zainmerchan) Julia-1 for Apple silicon. Runs Supersonic Labs' Julia-1 decision model on the Mac's GPU with MLX through the julia-mlx runtime: the same answers as the official PyTorch runtime on its published evaluations, 5–14× faster on the same Mac. The upstream model repository is SupersonicLabs/Julia-1. The files in this repository are Supersonic Labs' Julia-1 checkpoint, unchanged (model.safetensors SHA-256 df853bf7fe424420011f3d0c47a05d7341aa9eefa7fb9f203ea4aada4ad95b72). The runtime maps it into MLX directly, so no converted copy is needed. Precision (dtype="float16") and embedding placement are load-time options rather than separate files. This is an independent project, not affiliated with or… Open weights apache-2.0 144M parameters mlx [View model](https://savrn.com/models/julia-1-mlx) ## Le Van Huy [All models and datasets](https://savrn.com/model-publishers/huyleit) ## Versions - [0d954dabf3fa](https://savrn.com/models/phobert-emotion-social/versions/0d954dabf3fa) · current 2026-10-09 ## Explore More - [All text classification models](https://savrn.com/models/tasks/text-classification) - [All models under mit](https://savrn.com/models/licenses/mit) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-09. - [Hugging Face record](https://huggingface.co/huyleit/phobert-emotion-social) - [How the hub is built](https://savrn.com/model-hub/methodology)