HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

Qwen3-Embedding-8B

Qwen/Qwen3-Embedding-8B

kind
original
quant
bf16 (16-bit)
family
qwen
modality
feature-extraction
root model
Qwen/Qwen3-8B-Base
provider
Qwen
license
apache-2.0
language
—
context
32k
library
—
vRAM (GB)
8
size on disk
—
downloads
3,331,235
likes
758
created
—
last modified
2025-07-07T09:02:21+00:00 (2025)
huggingface
https://huggingface.co/Qwen/Qwen3-Embedding-8B
tags
sentence-transformerssafetensorsqwen3text-generationtransformerssentence-similarityfeature-extractiontext-embeddings-inferencearxiv:2506.05176base_model:Qwen/Qwen3-8B-Basebase_model:finetune:Qwen/Qwen3-8B-Baselicense:apache-2.0endpoints_compatibledeploy:azureregion:us

Benchmark scores (0)

no scores extracted for this variant.

Discussions (28)

titleauthorstatuscommentstype
Missing lm_head weights prevent loading in vLLM/Transformers (RuntimeError)rDevDoclosed0disc
NaN outputs when running in torch.float16 with text containing specific tokens (e.g., "import")Hank-Panopen0disc
how to use multi-gpu for inference?hh1212open0disc
如果想给embedding模型添加lm_heads该怎么做?linln1open0disc
Update README.mdRealWorgaopen0disc
Update README.mdaari1995open0disc
How to make the model generate a 32-dimensional vectortalenm01open0disc
Issue with Token 474 Causing NaN Embeddings in 8B ModelSernaryonopen0disc
dimensions not working it always generates 4096khanakiaopen0disc
Training for classification tasknastyafairyproopen0disc
Update `README.md` to use TEI v1.7 insteadalvarobarttopen0disc
Is `tie_word_embeddings: false` correct for this model?depasqualeopen0disc
Update README.md with TEI supportalvarobarttmerged0disc
sentence-transformers with 8Balmugaboopen0disc
MRL Performance losspsych0v0yageropen0disc
ONNX Conversionshuttieopen0disc
Downloading issueElijahhhhhopen0disc
The invoke results seems quite bad, did I code something wrong?ShuyangShiopen0disc
The result is completely different from the example.lightingfishopen0disc
Add missing `torch_dtype` in commented snippet in `README.md`alvarobarttopen0disc
Improve model cardnielsrclosed0disc
Training supportedtastelikefeetopen0disc
как уменьшить размер journalctl linuxwoldemarusopen0disc
add use case for vllm & modify Citationzyznullmerged0disc
希望为TEI推理添加支持forseasonopen0disc
不知道能否支持少数民族语言dojeeeopen0disc
Add last token poolingmichaelfeilopen0disc
Automatically add EOS via Tokenizer, integrate Sentence Transformerstomaarsenmerged0disc
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.