Skip to content
Language models

tiny-embed-27m

a compact and efficient Chinese text embedding model

2 min readupdated

The tiny-embed-27m model converts Chinese text into dense vector representations suited for semantic search, classification, clustering, and retrieval. Distilled from larger open models (including the BAAI bge family), it packs 27.6 million parameters with an output dimension of 512. Its standout architectural feature is recursive depth, where 4 unique layers are applied 16 times. The model needs no special query prefixes and outperforms similarly sized alternatives in benchmarks.

Key features

  • High performance: Scores an average of 0.5963 on the MTEB benchmark (Chinese suite, 31 tasks), outperforming bge-small-zh-v1.5 (24M).
  • Recursive architecture: 4 unique layers repeated 16 times keep the parameter count low while maintaining high inference speed.
  • Ready-to-use pipeline: Uses a standard Transformer → Pooling(mean) → Normalize setup with no extra projection layers.
  • No prefixes needed: Queries and documents take the exact same raw text without additional instructions.

What you can do

  • Semantic search: Plug the model into search engines to match short queries against long Chinese text documents.
  • Clustering and classification: Use vector representations to group similar texts and documents.
  • Quick integration: Run the model via the popular sentence-transformers library with trust_remote_code=True enabled.

Example

from sentence_transformers import SentenceTransformer

sentences_1 = ["样例数据-1", "样例数据-2"]
sentences_2 = ["样例数据-3", "样例数据-4"]

model = SentenceTransformer("dynasty-ai/tiny-embed-27m", trust_remote_code=True)
embeddings_1 = model.encode(sentences_1, normalize_embeddings=True)
embeddings_2 = model.encode(sentences_2, normalize_embeddings=True)
similarity = embeddings_1 @ embeddings_2.T
print(similarity)
Read moreDetails
• Metrics: Average MTEB(cmn, v1) score is 59.63 (compared to 57.34 for bge-small-zh-v1.5).
• Limitations: Maximum sequence length is 512 tokens.
• License: Released under the Apache-2.0 license, allowing commercial use.
• Tokenizer: Uses the standard Chinese BERT vocabulary of 21,128 tokens.
Discuss Telegram 1 23

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.7 GB
  • FP8 / INT8recommendedweights 0.1 GB0.7 GB
  • 4-bitweights 0.0 GB0.6 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

A ready script for the machine we advise above: paste it and the model downloads and starts.

  1. 1. Rent A30

    $0.26 an hour with the coupon. Pick the Ubuntu image when you create the machine.

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # dynasty-ai/tiny-embed-27m — A30, FP8 / INT8
    # NVIDIA driver: CUDA 13 needs 580+. If older or missing, the current nvidia-open is installed (keep yours: KEEP_DRIVER=1)
    DRV=$(nvidia-smi --query-gpu=driver_version --format=csv,noheader 2>/dev/null | head -1 | cut -d. -f1)
    case "$DRV" in ""|*[!0-9]*) DRV=0 ;; esac
    if [ "$DRV" -lt 580 ] && [ -z "${KEEP_DRIVER:-}" ]; then
      . /etc/os-release && \
      curl -fsSLO "https://developer.download.nvidia.com/compute/cuda/repos/ubuntu${VERSION_ID//./}/x86_64/cuda-keyring_1.1-1_all.deb" && \
      sudo dpkg -i cuda-keyring_1.1-1_all.deb && rm -f cuda-keyring_1.1-1_all.deb && \
      sudo apt-get update && \
      sudo DEBIAN_FRONTEND=noninteractive apt-get install -y nvidia-open || \
      { echo "Could not install the NVIDIA driver: see the error above. To keep the current one, run with KEEP_DRIVER=1."; exit 1; }
      echo "NVIDIA driver updated. Reboot the machine (sudo reboot) and run this same script again."
      exit 0
    fi
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv venv ~/vllm --python 3.12 && source ~/vllm/bin/activate
    uv pip install vllm
    vllm serve dynasty-ai/tiny-embed-27m --max-model-len 8192 --quantization fp8
    Add more LoRAs to tiny-embed-27m
  3. 3. Use it

    The server listens on port 8000 and speaks the OpenAI API: http://localhost:8000/v1 on the machine. From your computer: ssh -L 8000:localhost:8000 user@machine-IP.

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

  • MTP (multi-token prediction)0.34.0from November 16, 2026new release

    1.5–2× generation speed

    If the model has MTP layers (DeepSeek-V3/R1, Qwen3-Next, GLM, MiMo and others), it predicts several tokens per step itself.

    How to enable: vLLM / SGLang: --speculative-config '{"method": "mtp", "num_speculative_tokens": 1}'. Check the model card.

    Earlier generations (2)
    • EAGLE-3 — Speculative decoding with a small draft head. For models without their own MTP layers.
    • EAGLE-2 / Medusa — Earlier draft-head generations. Still work if that is all you find.
  • FP8 KV cache0.34.0from November 16, 2026new release

    2× more context in the same memory

    Keeps the attention cache in 8 bits, so long contexts fit where they did not.

    How to enable: vLLM: --kv-cache-dtype fp8; llama.cpp: -ctk q8_0 -ctv q8_0.

  • vLLM / SGLang0.34.0from November 16, 2026new release

    many times more tokens/s on a server

    Server engines: batching, FlashAttention, FP8 and speculative decoding out of the box.

    How to enable: vllm serve <model> — an OpenAI-compatible API.

    Earlier generations (1)
    • llama.cpp — GGUF models on any GPU, Mac or even CPU; some layers can stay in RAM.

Model family

New posts about versions of tiny-embed-27m (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

dynasty-ai/tiny-embed-27m
Author
dynasty-ai
Category
Language models
Released
Posted

On NeyroLearn

Hugging Face

231

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments