tag stringlengths 1 2.83k | kind stringclasses 25
values | model_count int64 1 3.24M | downloads_30d int64 0 2.7B | downloads_all_time int64 0 79.1B | top_pipeline_tag stringclasses 56
values | top_pipeline_share float64 0 1 ⌀ |
|---|---|---|---|---|---|---|
region:us | region | 3,235,409 | 2,696,136,571 | 79,090,483,718 | text-generation | 0.135 |
safetensors | library | 1,464,092 | 1,712,522,264 | 58,591,044,843 | text-generation | 0.233 |
endpoints_compatible | deploy | 1,077,383 | 1,871,216,814 | 63,430,176,374 | text-generation | 0.318 |
transformers | library | 994,685 | 1,685,965,203 | 65,264,886,012 | text-generation | 0.365 |
license:apache-2.0 | license | 594,793 | 1,506,490,278 | 42,170,685,000 | text-generation | 0.215 |
text-generation | pipeline_tag | 441,654 | 471,603,560 | 10,986,787,958 | text-generation | 0.987 |
en | language | 426,014 | 1,245,422,545 | 40,972,300,994 | text-generation | 0.302 |
conversational | freeform | 417,358 | 730,491,308 | 10,664,715,884 | text-generation | 0.623 |
arxiv:1910.09700 | arxiv | 398,406 | 39,224,565 | 2,215,112,252 | text-generation | 0.29 |
text-generation-inference | freeform | 379,461 | 283,156,239 | 9,558,487,448 | text-generation | 0.719 |
generated_from_trainer | freeform | 283,161 | 31,835,890 | 991,218,581 | text-generation | 0.158 |
peft | library | 276,184 | 2,388,258 | 52,524,044 | text-generation | 0.164 |
pytorch | library | 263,305 | 1,091,246,601 | 55,594,139,462 | text-generation | 0.218 |
llama | freeform | 229,641 | 58,367,070 | 2,699,553,613 | text-generation | 0.466 |
gguf | library | 221,881 | 424,281,233 | 3,094,998,211 | text-generation | 0.194 |
license:mit | license | 215,374 | 580,199,676 | 15,500,296,498 | text-generation | 0.123 |
tensorboard | library | 184,947 | 25,963,282 | 960,277,927 | text-classification | 0.161 |
trl | freeform | 163,876 | 13,562,211 | 300,119,126 | text-generation | 0.401 |
lora | freeform | 148,200 | 27,789,298 | 192,165,669 | text-to-image | 0.475 |
unsloth | freeform | 147,598 | 97,138,755 | 761,550,672 | text-generation | 0.307 |
license:other | license | 131,013 | 250,661,120 | 2,812,854,983 | text-to-image | 0.325 |
text-classification | pipeline_tag | 129,541 | 141,530,708 | 5,371,702,450 | text-classification | 0.99 |
diffusers | library | 124,481 | 70,655,934 | 1,610,647,606 | text-to-image | 0.836 |
model-index | freeform | 118,420 | 241,530,452 | 8,046,750,579 | reinforcement-learning | 0.522 |
qwen2 | freeform | 114,400 | 51,954,038 | 1,699,457,178 | text-generation | 0.428 |
gensyn | freeform | 110,677 | 392,712 | 46,759,364 | text-generation | 0.09 |
text-to-image | pipeline_tag | 105,062 | 71,449,062 | 1,090,970,577 | text-to-image | 0.994 |
minecraft | freeform | 98,797 | 77,973 | 1,174,839 | text-generation | 0 |
blockassist | freeform | 98,587 | 52 | 348 | question-answering | 0 |
gensyn-blockassist | freeform | 98,584 | 52 | 347 | null | null |
arxiv:2504.07091 | arxiv | 98,578 | 52 | 347 | null | null |
bert | freeform | 97,681 | 484,346,898 | 17,121,049,791 | text-classification | 0.471 |
text-embeddings-inference | freeform | 95,245 | 562,104,019 | 14,097,558,470 | text-classification | 0.749 |
sft | freeform | 85,849 | 6,317,708 | 72,938,059 | text-generation | 0.54 |
reinforcement-learning | pipeline_tag | 84,716 | 2,225,922 | 18,097,251 | reinforcement-learning | 0.954 |
qwen3 | freeform | 76,411 | 115,921,090 | 1,467,409,188 | text-generation | 0.475 |
onnx | library | 63,376 | 660,198,361 | 21,699,761,644 | reinforcement-learning | 0.22 |
custom_code | other | 60,006 | 109,416,231 | 2,472,746,673 | text-generation | 0.318 |
mistral | freeform | 55,269 | 11,052,703 | 821,969,847 | text-generation | 0.554 |
image-text-to-text | pipeline_tag | 54,708 | 373,086,262 | 4,128,207,911 | image-text-to-text | 0.86 |
text2text-generation | freeform | 53,733 | 44,789,960 | 2,327,239,672 | translation | 0.13 |
4-bit | other | 52,217 | 51,849,519 | 1,263,542,809 | text-generation | 0.551 |
axolotl | freeform | 50,028 | 6,219,834 | 117,221,624 | text-generation | 0.086 |
deep-reinforcement-learning | freeform | 49,160 | 832,578 | 10,137,137 | reinforcement-learning | 1 |
gpt2 | freeform | 46,116 | 41,143,589 | 1,521,801,739 | text-generation | 0.734 |
merge | other | 45,409 | 8,845,116 | 151,567,766 | text-generation | 0.475 |
feature-extraction | pipeline_tag | 45,319 | 506,404,652 | 13,805,478,855 | feature-extraction | 0.491 |
distilbert | freeform | 44,236 | 20,031,314 | 1,924,995,948 | text-classification | 0.599 |
base_model:black-forest-labs/FLUX.1-dev | base_model | 43,427 | 1,599,963 | 48,595,618 | text-to-image | 0.988 |
bitsandbytes | freeform | 43,151 | 5,746,729 | 154,359,671 | text-generation | 0.336 |
base_model:adapter:black-forest-labs/FLUX.1-dev | base_model | 42,686 | 1,014,992 | 35,145,749 | text-to-image | 0.997 |
robotics | pipeline_tag | 42,132 | 2,343,645 | 35,042,562 | robotics | 0.966 |
license:openrail | license | 41,847 | 3,966,883 | 89,399,467 | text-to-image | 0.027 |
automatic-speech-recognition | pipeline_tag | 40,039 | 241,842,162 | 4,711,142,960 | automatic-speech-recognition | 0.991 |
zh | language | 39,157 | 292,042,506 | 6,454,306,056 | text-generation | 0.335 |
mergekit | freeform | 38,803 | 5,935,804 | 114,012,791 | text-generation | 0.487 |
imatrix | freeform | 37,963 | 117,471,129 | 734,217,828 | text-generation | 0.159 |
8-bit | other | 37,333 | 52,919,290 | 1,007,313,313 | text-generation | 0.306 |
flux | freeform | 36,895 | 11,374,513 | 141,177,247 | text-to-image | 0.977 |
roberta | freeform | 35,853 | 41,971,709 | 3,798,100,360 | text-classification | 0.473 |
t5 | freeform | 33,619 | 95,800,959 | 2,569,696,247 | text-generation | 0.036 |
license:creativeml-openrail-m | license | 33,442 | 6,980,318 | 476,888,140 | text-to-image | 0.575 |
base_model:Qwen/Qwen1.5-0.5B | base_model | 32,534 | 69,126 | 1,158,219 | text-generation | 0.005 |
base_model:adapter:Qwen/Qwen1.5-0.5B | base_model | 32,473 | 48,068 | 597,626 | text-generation | 0.004 |
lerobot | freeform | 31,787 | 724,499 | 4,572,437 | robotics | 0.995 |
token-classification | pipeline_tag | 31,726 | 52,544,255 | 1,819,263,326 | token-classification | 0.994 |
stable-diffusion | freeform | 31,687 | 16,207,882 | 816,726,330 | text-to-image | 0.969 |
image-classification | pipeline_tag | 30,825 | 69,185,769 | 5,895,349,169 | image-classification | 0.99 |
base_model:Qwen/Qwen1.5-1.8B | base_model | 30,628 | 49,579 | 561,591 | text-generation | 0.005 |
base_model:adapter:Qwen/Qwen1.5-1.8B | base_model | 30,581 | 46,546 | 486,164 | text-generation | 0.004 |
custom-implementation | freeform | 30,064 | 2,391 | 25,902 | reinforcement-learning | 1 |
mlx | library | 28,987 | 47,511,656 | 481,333,864 | text-generation | 0.455 |
stable-baselines3 | library | 28,350 | 317,430 | 5,898,017 | reinforcement-learning | 1 |
sentence-transformers | library | 27,572 | 546,526,296 | 13,214,235,918 | sentence-similarity | 0.723 |
whisper | freeform | 27,103 | 100,941,400 | 897,990,770 | automatic-speech-recognition | 0.766 |
llama-cpp | freeform | 26,459 | 19,660,617 | 87,983,803 | text-generation | 0.377 |
es | language | 26,259 | 264,451,982 | 7,403,130,045 | text-generation | 0.358 |
fr | language | 26,168 | 287,340,198 | 8,075,908,505 | text-generation | 0.359 |
grpo | freeform | 26,087 | 1,512,391 | 10,409,163 | text-generation | 0.732 |
license:cc-by-nc-4.0 | license | 25,189 | 40,970,057 | 1,028,225,253 | text-generation | 0.21 |
xlm-roberta | freeform | 24,425 | 108,767,200 | 3,177,770,008 | text-classification | 0.3 |
base_model:google/gemma-2b | base_model | 24,037 | 116,739 | 3,581,258 | text-generation | 0.013 |
base_model:adapter:google/gemma-2b | base_model | 23,700 | 99,271 | 3,104,475 | text-generation | 0.007 |
de | language | 23,408 | 278,517,658 | 7,760,021,331 | text-generation | 0.358 |
sentence-similarity | pipeline_tag | 22,948 | 445,334,021 | 11,493,707,053 | sentence-similarity | 0.911 |
qwen3_5 | freeform | 22,872 | 105,559,294 | 719,725,807 | image-text-to-text | 0.378 |
replicate | freeform | 22,666 | 243,585 | 3,384,501 | text-to-image | 0.99 |
ja | language | 22,332 | 244,747,487 | 6,212,888,934 | text-generation | 0.316 |
quantized | freeform | 22,025 | 142,377,902 | 966,046,284 | text-generation | 0.427 |
gguf-my-repo | freeform | 21,435 | 2,673,793 | 34,320,830 | text-generation | 0.335 |
ko | language | 21,084 | 216,811,886 | 5,653,691,847 | text-generation | 0.354 |
translation | pipeline_tag | 19,841 | 32,292,005 | 1,132,979,576 | translation | 0.962 |
pt | language | 19,202 | 232,202,555 | 6,821,647,195 | text-generation | 0.358 |
fill-mask | pipeline_tag | 19,016 | 132,565,974 | 12,021,271,213 | fill-mask | 0.976 |
it | language | 18,665 | 231,928,508 | 6,670,554,490 | text-generation | 0.389 |
stable-diffusion-xl | freeform | 18,004 | 4,208,041 | 104,662,250 | text-to-image | 0.982 |
license:openrail++ | license | 17,737 | 10,840,054 | 294,510,245 | text-to-image | 0.899 |
not-for-all-audiences | freeform | 17,700 | 11,069,040 | 1,543,773,487 | text-to-image | 0.296 |
tf | library | 17,503 | 489,336,781 | 27,961,769,775 | text-classification | 0.191 |
q-learning | freeform | 17,076 | 519 | 2,029 | reinforcement-learning | 1 |
Hugging Face Model Tags
Every distinct tag on every model on the Hugging Face Hub, deduplicated, with how many models use it. Snapshot of 2026-11-15: 3,239,126 models, 456,200 distinct tags, of which 135,741 are freeform (author-chosen) tags; 73,158 freeform tags are used by exactly one model.
Built for Taxonomy of AI: authors' own tags are a crowd-sourced classification of models, noisy but rich, and this makes the whole vocabulary queryable.
Configs
tags: one row per distinct tag (exact string, case-sensitive).
| Column | Meaning |
|---|---|
tag |
the tag as it appears in a model's tags list |
kind |
the Hub's official tag type if the tag is in GET /api/models-tags-by-type (language, library, license, region, pipeline_tag, dataset, deploy, other, bucket); else the prefix before : when 50+ distinct tags share it (base_model, arxiv, doi, ...); else freeform |
model_count |
models carrying the tag |
downloads_30d, downloads_all_time |
summed over those models |
top_pipeline_tag |
the pipeline_tag most common among those models (models without one are ignored) |
top_pipeline_share |
models with that pipeline tag / model_count (low when most of the tag's models have no pipeline tag) |
freeform_folded: freeform tags merged across case and separators, so voice-conversion, Voice Conversion,
voice_conversion and VoiceConversion are one row. Columns: key (lowercased, letters and digits only),
most_common (the most-used spelling), model_count (sum over variants; a model tagged with two variants counts
twice), n_variants, variants (spellings, most used first).
tag_groups: all tags deduplicated into groups (added 2026-11-16; 439,946 groups). tag names the group: the
official tag if one is a member, else the most-used spelling. kind is that tag's kind. model_count, downloads_30d,
downloads_all_time are summed over members (a model tagged with two spellings counts twice). n_variants,
variants (tag, kind, model_count, most used first), merge: none, fold (case and separators), official
(a freeform spelling of an official tag joined it, e.g. 13 spellings of text-to-speech), stem (plural and
-ing/-ed forms, e.g. fine-tuned / finetune / fine_tuning). Model names with digits aren't stemmed; ID kinds
(base_model, dataset, arxiv, ...) aren't merged. Code: scripts/dedup_hub_tags.py. Browse it in the
HF Tag Explorer Space.
kinds: distinct tags and total taggings per kind.
Source and method
- Source:
cfahlgren1/hub-statsmodels.parquet(updated 2026-11-15T13:41Z), which mirrors the Hub's model API. Onlyid,tags,pipeline_tagand download columns are read. Official tag types fromhttps://huggingface.co/api/models-tags-by-typeon the same day. - Code:
scripts/build_hub_tags.py(DuckDB). Run on a Google Colab CPU VM in about a minute. - Querying the Hub by tag: the website's
?other=<tag>filter corresponds toGET /api/models?filter=<tag>(exact, case-sensitive). The API silently ignores another=parameter.
New snapshots go in snapshots/<date>/; the configs point at the latest one.
- Downloads last month
- 145