LF

LFM2.5 Encoder 350M

LfmGGUF by LiquidAIUpdated 2026-10-02
Download on Hugging Face
Parameters354M
Context125K
Layers16
Architecturelfm2
Smallest quant0.2 GB
Quick answer

LFM2.5 Encoder 350M needs about 1.4 GB of memory at Q8_0 (a 0.4 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 3 quants, from 0.2 GB (Q4_0) to 0.7 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 2.1 GB.

LFM2.5-Encoder 350M – a multilingual bidirectional encoder

LFM2.5-Encoder 350M is a 0.354 billion‑parameter masked language model with 8k context. It supports 15 languages and is designed for fine‑tuning into classification, token classification, retrieval, reranking, and semantic similarity tasks. The model runs efficiently on‑device and in the browser.

Good at

  • Multilingual encoding
  • Bidirectional attention
  • Fine‑tuning for downstream tasks

Limits

  • No vision support
  • License is other
  • No mixture of experts

Best for

  • Text classification
  • Semantic similarity
  • Retrieval ranking

Who it’s for. Developers building multilingual NLP pipelines who need a lightweight encoder that can run on CPUs or WebGPU. It is suitable for on‑device inference and fine‑tuning into task‑specific models.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_00.2 GB 1.3 GB
Fits Get file →
Q8_00.4 GB 1.4 GB
Fits Get file →
F160.7 GB 1.7 GB
Fits Get file →

For 8 GB VRAM, Q4_0 (0.22 GB) uses about 1.3 GB at 8k context. Q8_0 (0.38 GB) uses about 1.4 GB. For higher precision, F16 (0.71 GB) uses about 1.7 GB. Choose the smallest quant that meets your accuracy needs.

What GPU do you need for LFM2.5 Encoder 350M?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is F16 (0.7 GB file, about 1.7 GB in use).

Which LFM2.5 Encoder 350M quant should you choose?

Q8_0 (0.4 GB) is the middle option in this repository. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

If memory is tight, the smallest file is Q4_0 at 0.2 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturelfm2
Parameters354M
Layers16
Trained context128,000 tokens
KV cache (fp16)64 KB per token

Source

Quantized byLiquidAI
FamilyLfm
Licenseother
Downloads (HF)1,511
Last updated2026-10-02

Frequently asked questions about LFM2.5 Encoder 350M

How much VRAM does LFM2.5 Encoder 350M need?

About 1.4 GB at Q8_0 with an 8K context window. That is the 0.4 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run LFM2.5 Encoder 350M on an 8 GB GPU?

Yes, with F16: the 0.7 GB file needs about 1.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of LFM2.5 Encoder 350M GGUF?

It depends on the quant: Q4_0 is 0.2 GB, Q8_0 is 0.4 GB, F16 is 0.7 GB.

What context length does LFM2.5 Encoder 350M support?

The GGUF header lists a trained context length of 128,000 tokens (125K). The KV cache grows with the context you use: about 0.5 GB at 8K and 2.1 GB at 32K.

Where can I download LFM2.5 Encoder 350M GGUF?

The files are on Hugging Face in the LiquidAI/LFM2.5-Encoder-350M-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does LFM2.5 Encoder 350M use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run LFM2.5-Encoder 350M on 8 GB VRAM?

Yes. The Q4_0 quant uses about 1.3 GB of VRAM at 8k context, well within an 8 GB budget.

Does this model support vision tasks?

No. It is a text‑only encoder with no vision capabilities.

What languages does it support?

It supports 15 languages: English, German, Spanish, French, Italian, Dutch, Polish, Portuguese, Arabic, Hindi, Japanese, Russian, Turkish, Vietnamese, Chinese.

Other GGUFs of LFM2.5-Encoder-350M

Similar models

Browse all →

More from LiquidAI

All by LiquidAI →

More Lfm models

All Lfm →

Other tiny models

All tiny models →