SE

SeaLLM-7B-v2

SeallmGGUF by mradermacherUpdated 2024-08-02
Download on Hugging Face
Parameters7.38B
Context32K
Layers32
Architecturellama
Smallest quant1.7 GB
Quick answer

SeaLLM-7B-v2 needs about 6.0 GB of memory at Q4_K_M (a 4.5 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 21 quants, from 1.7 GB (IQ1_S) to 6.1 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.

SeaLLM-7B-v2 – 7.38B multilingual LLM for Southeast Asian languages

SeaLLM-7B-v2 is a 7.38B parameter model trained on 32768 context tokens. It supports multiple Southeast Asian languages and offers strong zero‑shot reasoning performance. The model is available in several GGUF quantizations.

Good at

  • Multilingual reasoning
  • Zero‑shot math and commonsense
  • Instruction following

Limits

  • No vision support
  • License is other
  • Requires seallm.preset.json for chat template

Best for

  • Local inference on 8‑GB VRAM
  • Low‑latency chat applications
  • Educational language tools

Who it’s for. Developers and researchers needing a compact multilingual LLM for Southeast Asian languages. The model is suitable for on‑device inference and can be run with llama.cpp or compatible runtimes.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S1.7 GB 3.3 GB
Fits Get file →
IQ1_M1.8 GB 3.4 GB
Fits Get file →
IQ2_XXS2.1 GB 3.6 GB
Fits Get file →
IQ2_XS2.3 GB 3.8 GB
Fits Get file →
IQ2_S2.4 GB 4.0 GB
Fits Get file →
IQ2_M2.6 GB 4.1 GB
Fits Get file →
Q2_K2.8 GB 4.4 GB
Fits Get file →
IQ3_XXS2.9 GB 4.5 GB
Fits Get file →
IQ3_XS3.1 GB 4.7 GB
Fits Get file →
Q3_K_S3.2 GB 4.8 GB
Fits Get file →
IQ3_S3.3 GB 4.8 GB
Fits Get file →
IQ3_M3.4 GB 4.9 GB
Fits Get file →
Q3_K_M3.6 GB 5.2 GB
Fits Get file →
Q3_K_L3.9 GB 5.5 GB
Fits Get file →
IQ4_XS4.0 GB 5.6 GB
Fits Get file →
Q4_04.2 GB 5.8 GB
Fits Get file →
Q4_K_S4.2 GB 5.8 GB
Fits Get file →
Q4_K_M4.5 GB 6.0 GB
Fits Get file →
Q5_K_S5.1 GB 6.7 GB
Fits Get file →
Q5_K_M5.2 GB 6.8 GB
Fits Get file →
Q6_K6.1 GB 7.6 GB
Tight Get file →

For 8‑GB VRAM, use Q4_0 (file size 4.22 GB, memory 5.8 GB at 8k context). If 12‑GB VRAM is available, Q5_K_S (5.1 GB, 6.7 GB) or Q5_K_M (5.23 GB, 6.8 GB) are good choices. For 16‑GB VRAM, Q6_K (6.05 GB, 7.6 GB) provides higher quality.

What GPU do you need for SeaLLM-7B-v2?

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 IQ2_S (2.4 GB file, about 4.0 GB in use).
  • 6 GB VRAM: the best quant that fits is Q4_K_M (4.5 GB file, about 6.0 GB in use).
  • 8 GB VRAM: the best quant that fits is Q6_K (6.1 GB file, about 7.6 GB in use).

Which SeaLLM-7B-v2 quant should you choose?

Q4_K_M is the usual starting point for SeaLLM-7B-v2: it balances file size and quality well. Here it is a 4.5 GB file.

If memory is tight, the smallest file is IQ1_S at 1.7 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters7.38B
Layers32
Trained context32,768 tokens
KV cache (fp16)128 KB per token

Source

Quantized bymradermacher
FamilySeallm
Licenseother
Downloads (HF)1,710
Last updated2024-08-02

Frequently asked questions about SeaLLM-7B-v2

How much VRAM does SeaLLM-7B-v2 need?

About 6.0 GB at Q4_K_M with an 8K context window. That is the 4.5 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 SeaLLM-7B-v2 on an 8 GB GPU?

Yes, with Q6_K: the 6.1 GB file needs about 7.6 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of SeaLLM-7B-v2 GGUF?

It depends on the quant: IQ1_S is 1.7 GB, IQ1_M is 1.8 GB, IQ2_XXS is 2.1 GB, IQ2_XS is 2.3 GB, IQ2_S is 2.4 GB, IQ2_M is 2.6 GB and 15 more.

What context length does SeaLLM-7B-v2 support?

The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.

Where can I download SeaLLM-7B-v2 GGUF?

The files are on Hugging Face in the mradermacher/SeaLLM-7B-v2-i1-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 SeaLLM-7B-v2 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 SeaLLM-7B-v2 on 8 GB VRAM?

Yes, the Q4_0 quant uses about 5.8 GB of VRAM at 8k context, which fits within an 8 GB budget.

Does SeaLLM-7B-v2 support vision tasks?

No, the model does not include vision capabilities.

What license applies to SeaLLM-7B-v2?

The model is released under an "other" license, so check the repository for usage terms.

Other GGUFs of SeaLLM-7B-v2

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Seallm models

All Seallm →

Other small models

All small models →