LFM2.5 ColBERT 350M needs about 1.3 GB of memory at Q4_K_M (a 0.2 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 7 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-ColBERT-350M: Multilingual late‑interaction retriever
LFM2.5-ColBERT-350M is a 350M‑parameter late‑interaction model that stores one vector per token. It supports 11 languages and is designed for fast, accurate retrieval in multilingual RAG pipelines.
Good at
- Multilingual retrieval
- Late‑interaction MaxSim matching
- Fast inference on small hardware
Limits
- No vision support
- Requires at least 8k context for memory estimates
- License is other
Best for
- E‑commerce semantic search
- FAQ and support knowledge bases
- On‑device semantic search
Who it’s for. Developers building multilingual RAG systems or semantic search engines that need high accuracy across many languages. The model is suitable for both cloud and on‑device deployments where memory is limited.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_0 | 0.2 GB | 1.3 GB | Fits | Get file → | |
| Q4_K_M | 0.2 GB | 1.3 GB | Fits | Get file → | |
| Q5_K_M | 0.3 GB | 1.3 GB | Fits | Get file → | |
| Q6_K | 0.3 GB | 1.3 GB | Fits | Get file → | |
| Q8_0 | 0.4 GB | 1.4 GB | Fits | Get file → | |
| BF16 | 0.7 GB | 1.7 GB | Fits | Get file → | |
| F16 | 0.7 GB | 1.7 GB | Fits | Get file → |
For 8 GB VRAM, Q4_0 or Q4_K_M (0.22–0.23 GB) use about 1.3 GB memory at 8k context. If 16 GB VRAM is available, Q5_K_M or Q6_K (0.26–0.29 GB) remain within 1.3 GB. Larger BF16 or F16 (0.71 GB) require more memory and are best for high‑end GPUs.
What GPU do you need for LFM2.5 ColBERT 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 BF16 (0.7 GB file, about 1.7 GB in use).
Which LFM2.5 ColBERT 350M quant should you choose?
Q4_K_M is the usual starting point for LFM2.5 ColBERT 350M: it balances file size and quality well. Here it is a 0.2 GB file.
The Q8_0 file is 0.4 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
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
Source
Frequently asked questions about LFM2.5 ColBERT 350M
How much VRAM does LFM2.5 ColBERT 350M need?
About 1.3 GB at Q4_K_M with an 8K context window. That is the 0.2 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 ColBERT 350M on an 8 GB GPU?
Yes, with BF16: 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 ColBERT 350M GGUF?
It depends on the quant: Q4_0 is 0.2 GB, Q4_K_M is 0.2 GB, Q5_K_M is 0.3 GB, Q6_K is 0.3 GB, Q8_0 is 0.4 GB, BF16 is 0.7 GB and 1 more.
What context length does LFM2.5 ColBERT 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 ColBERT 350M GGUF?
The files are on Hugging Face in the LiquidAI/LFM2.5-ColBERT-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 ColBERT 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-ColBERT-350M on 8 GB VRAM?
Yes, using Q4_0 or Q4_K_M quantizations will use about 1.3 GB memory at 8k context, fitting within 8 GB VRAM.
Does the model support image inputs?
No, the model only handles text embeddings and does not support vision.
What languages are supported?
English, Spanish, German, French, Italian, Portuguese, Arabic, Swedish, Norwegian, Japanese, Korean.