KO

KoSOLAR 10.7B merge v3.4 i1

KosolarGGUF by mradermacherUpdated 2024-12-11
Download on Hugging Face
Parameters10.8B
Context4K
Layers48
Architecturellama
Smallest quant2.4 GB
Quick answer

KoSOLAR 10.7B merge v3.4 i1 needs about 8.6 GB of memory at Q4_K_M (a 6.5 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 25 quants, from 2.4 GB (IQ1_S) to 8.9 GB (Q6_K).

KoSOLAR 10.7B merge v3.4 i1 – a 10.8B llama model

KoSOLAR 10.7B merge v3.4 i1 is a 10.8B parameter llama model merged with SLERP. It supports 4096 token context and is released under apache-2.0. The model is available in multiple GGUF quantizations.

Good at

  • Large context language tasks
  • Fine‑grained text generation
  • Multilingual support

Limits

  • No vision support
  • Requires 8.2‑11 GB VRAM for full precision
  • License is apache‑2.0

Best for

  • Chatbots with long context
  • Research prototyping
  • Content creation

Who it’s for. Developers and researchers needing a large llama model with flexible quant options. The model is suitable for local deployment on systems with sufficient VRAM.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S2.4 GB 3.7 GB
Fits Get file →
IQ1_M2.6 GB 3.9 GB
Fits Get file →
IQ2_XXS3.0 GB 4.3 GB
Fits Get file →
IQ2_XS3.3 GB 4.6 GB
Fits Get file →
IQ2_S3.4 GB 4.7 GB
Fits Get file →
IQ2_M3.7 GB 5.0 GB
Fits Get file →
Q2_K_S3.8 GB 5.1 GB
Fits Get file →
Q2_K4.0 GB 5.4 GB
Fits Get file →
IQ3_XXS4.2 GB 5.5 GB
Fits Get file →
IQ3_XS4.5 GB 5.8 GB
Fits Get file →
Q3_K_S4.7 GB 6.0 GB
Fits Get file →
IQ3_S4.7 GB 6.0 GB
Fits Get file →
IQ3_M4.9 GB 6.2 GB
Fits Get file →
Q3_K_M5.2 GB 6.5 GB
Fits Get file →
Q3_K_L5.7 GB 7.0 GB
Fits Get file →
IQ4_XS5.8 GB 7.1 GB
Fits Get file →
Q4_06.1 GB 7.4 GB
Tight Get file →
Q4_06.1 GB 7.4 GB
Tight Get file →
Q4_06.1 GB 7.4 GB
Tight Get file →
Q4_06.1 GB 7.5 GB
Tight Get file →
Q4_K_S6.2 GB 7.5 GB
Tight Get file →
Q4_K_M6.5 GB 7.8 GB
Tight Get file →
Q5_K_S7.5 GB 8.8 GB
Too big Get file →
Q5_K_M7.7 GB 9.0 GB
Too big Get file →
Q6_K8.9 GB 10.2 GB
Too big Get file →

For 8 GB VRAM, use Q4_0 (6.12 GB file, 8.2 GB memory). For 10 GB VRAM, Q5_K_M (7.65 GB file, 9.8 GB memory). For 12 GB VRAM, Q6_K (8.87 GB file, 11 GB memory). Estimates are approximate.

What GPU do you need for KoSOLAR 10.7B merge v3.4 i1?

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: no quant fits at an 8K context.
  • 6 GB VRAM: the best quant that fits is Q2_K_S (3.8 GB file, about 5.9 GB in use).
  • 8 GB VRAM: the best quant that fits is IQ4_XS (5.8 GB file, about 7.9 GB in use).
  • 12 GB VRAM: the best quant that fits is Q6_K (8.9 GB file, about 11.0 GB in use).

Which KoSOLAR 10.7B merge v3.4 i1 quant should you choose?

Q4_K_M is the usual starting point for KoSOLAR 10.7B merge v3.4 i1: it balances file size and quality well. Here it is a 6.5 GB file.

If memory is tight, the smallest file is IQ1_S at 2.4 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
Parameters10.8B
Layers48
Trained context4,096 tokens
KV cache (fp16)192 KB per token

Source

Quantized bymradermacher
FamilyKosolar
Licenseapache-2.0
Downloads (HF)800
Last updated2024-12-11

Frequently asked questions about KoSOLAR 10.7B merge v3.4 i1

How much VRAM does KoSOLAR 10.7B merge v3.4 i1 need?

About 8.6 GB at Q4_K_M with an 8K context window. That is the 6.5 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 KoSOLAR 10.7B merge v3.4 i1 on an 8 GB GPU?

Yes, with IQ4_XS: the 5.8 GB file needs about 7.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of KoSOLAR 10.7B merge v3.4 i1 GGUF?

It depends on the quant: IQ1_S is 2.4 GB, IQ1_M is 2.6 GB, IQ2_XXS is 3.0 GB, IQ2_XS is 3.3 GB, IQ2_S is 3.4 GB, IQ2_M is 3.7 GB and 19 more.

What context length does KoSOLAR 10.7B merge v3.4 i1 support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 1.6 GB at 8K and 0.8 GB at 4K.

Where can I download KoSOLAR 10.7B merge v3.4 i1 GGUF?

The files are on Hugging Face in the mradermacher/KoSOLAR-10.7B-merge-v3.4-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 KoSOLAR 10.7B merge v3.4 i1 use?

Hugging Face lists the license as apache-2.0. 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 this model on 8 GB VRAM?

Yes, use the Q4_0 quant which requires about 8.2 GB VRAM at 8k context.

Does this model support vision tasks?

No, it is a pure language model with no vision capabilities.

What license applies to this model?

The model is released under the apache-2.0 license.

Other GGUFs of KoSOLAR-10.7B-merge-v3.4

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Kosolar models

All Kosolar →

Other medium models

All medium models →