alt.freeweights.* · 29 sections · 8900 models listed (2868 shown under their base as variants) · collection

Runs in the browser

Models with an ONNX build small enough for a browser tab with WebGPU: about 1.33 GB of files or less (a 1.6 GB GPU buffer, file size x 1.2). Hand-picked models come first. Phone browsers allow about 0.36 GB, so only builds of 300 MB or less fit there.

Filters

Sections

alt.freeweights.*6279 qwen655 deepseek82 kimi34 glm182 minimax29 mimo11 gemma376 gpt-oss9 olmo575 nemotron109 phi66 mistral118 granite162 smollm49 minicpm106 llama254 lfm74 text1134 speech186 embed219 vision1225 image199 video73 music16 ocr1 datasets11 tools247 abliterated77 allternit

Collections

Runs on a 256 GB Mac3692 Small enough for a laptop2556 Embeddings and rerankers290 Speech276 GGUF for llama.cpp and Ollama694 Open datasets11 Runs on a phone450 Runs on a laptop864 Needs a workstation455 Runs in the browser338

Filter

Sectionvision152embed60text37speech32qwen16lfm14

Show 7 more

granite11smollm7olmo4nemotron2abliterated1gemma1llama1

Tasktext generation60sentence similarity30automatic speech recognition28feature extraction25image classification23object detection22

Show 22 more

text classification21token classification17fill mask13image text to text13zero shot image classification11image feature extraction9not tagged8text ranking8zero shot classification8image segmentation7question answering6zero shot object detection5text to speech4translation4image to text3keypoint detection3audio classification2mask generation2summarization2voice activity detection2depth estimation1time series forecasting1

License classPermissive319Custom terms19

Formatsafetensors324onnx78gguf6openvino5ggml4coreml2

Show 2 more

ctranslate21mlx1

SizeUnder 4 GB2954 to 40 GB4240 to 200 GB1

TermsNo extra terms319Requires accepting terms19

ParametersUnder 1B2921B to 3B31Not stated15

Librarysafetensors324transformers282PyTorch129ONNX78sentence-transformers62Transformers.js37

Show 9 more

JAX33timm11GGUF6OpenVINO5GGML4Core ML2CTranslate21MLX1open_clip1

LanguageEnglish196German57French56Japanese52Spanish51Chinese49

Show 110 more

Arabic42Italian40Korean39Portuguese36Russian36Hindi34Polish34Dutch33Multilingual31Turkish31Vietnamese31Hebrew29Swedish29Greek28Indonesian28Thai28Danish27Persian27Romanian27Swahili27Bulgarian26Finnish26Malay26Ukrainian26Hungarian25Georgian25Latvian25Marathi25Slovenian25Albanian25Urdu25Bangla24Catalan24Czech24Estonian24Croatian24Lithuanian24Slovak24Serbian24Azerbaijani23Armenian23Icelandic23Burmese23Telugu23Filipino23Galician22Gujarati22Kazakh22Macedonian22Mongolian22Tamil22Afrikaans21Welsh21Kannada21Malayalam21Belarusian20Basque20Khmer20Nepali20Punjabi20Latin18Norwegian18Uzbek18Lao17Sinhala17Breton16Bosnian16Haitian Creole16Malagasy16Yoruba16Amharic15Somali15Javanese14Sundanese14Tatar13Norwegian Nynorsk12Occitan12Pashto12Tajik12Assamese11Hausa11Kyrgyz11Sanskrit11Sindhi11Yiddish11Bashkir10Irish9Luxembourgish9Norwegian Bokmål8Tibetan7Faroese7Lingala7Māori7Maltese7Shona7Turkmen7Esperanto6Kurdish6Xhosa6Volapük5Western Frisian4Scottish Gaelic4Oromo4Odia4Quechua4Uyghur4Aragonese3Chechen3Chuvash3Ido3

AppsSGLang79vLLM79llama.cpp4LM Studio4Ollama4whisper.cpp4

Show 1 more

Jan1

On devicePhone338Laptop338In the browser338

Enginewebgpu338llama.cpp45mlx28

all 6279 Runs on a 256 GB Mac 3692 Small enough for a laptop 2556 Embeddings and rerankers 290 Speech 276 GGUF for llama.cpp and Ollama 694 Open datasets 11 Runs on a phone 450 Runs on a laptop 864 Needs a workstation 455 Runs in the browser 338 qwen 655 deepseek 82 kimi 34 glm 182 minimax 29 mimo 11 gemma 376 gpt-oss 9 olmo 575 nemotron 109 phi 66 mistral 118 granite 162 smollm 49 minicpm 106 llama 254 lfm 74 text 1134 speech 186 embed 219 vision 1225 image 199 video 73 music 16 ocr 1 datasets 11 tools 247 abliterated 77

Search models/

Sort

338 models

SubjectSizeParamsLicenseDate

grounding-dino-tiny-ONNXHosted (signed) From: onnx-community in alt.freeweights.vision2.29 GB—Permissive2026-11-15 mobilebert-finetuned-posHosted (signed) From: mrm8488 in alt.freeweights.vision296 MB25MPermissive2026-11-15 rubert-tinyHosted (signed) From: cointegrated in alt.freeweights.vision367 MB12MPermissive2026-11-15 dfine-small-cocoHosted (signed) From: ustc-community in alt.freeweights.vision41.5 MB10MPermissive2026-11-15 AI-Anime-Image-Detector-ViTHosted (signed) From: legekka in alt.freeweights.vision350 MB88MPermissive2026-11-15 MonadHosted (signed) From: PleIAs in alt.freeweights.text114 MB57MPermissive2026-11-15 kotoba-whisper-bilingual-v1.0Hosted (signed) From: kotoba-tech in alt.freeweights.speech4.54 GB756MPermissive2026-11-15

dictabert-ner-ONNXHosted (signed) From: onnx-community in alt.freeweights.text2.80 GB—Permissive2026-11-15 nougat-latex-baseHosted (signed) From: Norm in alt.freeweights.vision2.79 GB349MPermissive2026-11-15 dfine-medium-cocoHosted (signed) From: ustc-community in alt.freeweights.vision78.7 MB20MPermissive2026-11-15

← PreviousPage 3 of 4Next →

Slide, drag, or use the arrow keys.

No model matches these filters.Remove a filter or search for something shorter.

Every row is downloadable or links to its source. Press / to search, n and p to move in the list.

From: PekingU <huggingface.co/PekingU>Newsgroups: alt.freeweights.visionSubject: rtdetr_v2_r50vdDate: 2026-11-15Commit: 2824940Signed: 2DC28CC7B26A283B ✓

Object detection model with 43M parameters. Safetensors weights, 172 MB in all. Published by PekingU under Apache-2.0.

43M params172MBrtdetr_v2_r50vdsafetensorspermissive

FactsParams43MSize172 MBFormatssafetensorsTaskobject detectionLicenseApache-2.0From sourcepermissive

$ fw get PekingU/rtdetr_v2_r50vd
# manifest sha256 6e9ed3e71629… · 172 MB · 5 files
$ fw verify ./rtdetr_v2_r50vd PekingU/rtdetr_v2_r50vd

Open model page

Goal 1: 45 USD, 3 months x 15 USD; or a verified free VM0% fundedREADME: operating statementAllternit, the open source alternative