GR

Granite 3.0 3b A800M Instruct

GraniteGGUF by bartowskiUpdated 2024-10-21
Download on Hugging Face
Parameters3.37B
Context4K
Layers32
Architecturegranitemoe
Smallest quant1.3 GB
Quick answer

Granite 3.0 3b A800M Instruct needs about 3.1 GB of memory at Q4_K_M (a 2.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 23 quants, from 1.3 GB (Q2_K) to 6.8 GB (F16).

Granite 3.0 3b A800M Instruct – 3.37B parameter instruction model

Granite 3.0 3b A800M Instruct is a 3.37B parameter decoder‑only sparse Mixture of Experts model. It was finetuned from Granite‑3.0‑3B‑A800M‑Base‑4K using open source instruction datasets and internal synthetic data. The model supports 12 languages and is licensed under Apache 2.0.

Good at

  • Instruction following
  • Multilingual dialogue
  • Code and function‑calling tasks

Limits

  • Requires 4096 token context
  • No vision support
  • Large memory for higher quant levels

Best for

  • Building AI assistants
  • Retrieval‑augmented generation
  • Text classification and summarization

Who it’s for. Developers and researchers building instruction‑driven AI assistants or multilingual applications. The model is suitable for users who need a balance between performance and memory usage and who can run it on GPU or CPU with llama.cpp.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K1.3 GB 2.0 GB
Fits Get file →
Q2_K_L1.3 GB 2.1 GB
Fits Get file →
IQ3_XS1.4 GB 2.2 GB
Fits Get file →
Q3_K_S1.5 GB 2.3 GB
Fits Get file →
IQ3_M1.5 GB 2.3 GB
Fits Get file →
Q3_K_M1.6 GB 2.4 GB
Fits Get file →
Q3_K_L1.8 GB 2.5 GB
Fits Get file →
IQ4_XS1.8 GB 2.6 GB
Fits Get file →
Q3_K_XL1.8 GB 2.6 GB
Fits Get file →
Q4_01.9 GB 2.7 GB
Fits Get file →
Q4_01.9 GB 2.7 GB
Fits Get file →
Q4_01.9 GB 2.7 GB
Fits Get file →
Q4_01.9 GB 2.7 GB
Fits Get file →
Q4_K_S1.9 GB 2.7 GB
Fits Get file →
Q4_K_M2.1 GB 2.8 GB
Fits Get file →
Q4_K_L2.1 GB 2.9 GB
Fits Get file →
Q5_K_S2.3 GB 3.1 GB
Fits Get file →
Q5_K_M2.4 GB 3.2 GB
Fits Get file →
Q5_K_L2.5 GB 3.2 GB
Fits Get file →
Q6_K2.8 GB 3.5 GB
Fits Get file →
Q6_K_L2.8 GB 3.6 GB
Fits Get file →
Q8_03.6 GB 4.4 GB
Fits Get file →
F166.8 GB 7.5 GB
Tight Get file →

For 8 GB VRAM, use Q4_K_M (file size 2.06 GB, estimated 3.1 GB memory). For 12 GB VRAM, Q5_K_M (file size 2.41 GB, estimated 3.4 GB memory) is a good choice. For 16 GB VRAM, Q6_K (file size 2.78 GB, estimated 3.8 GB memory) offers higher quality.

What GPU do you need for Granite 3.0 3b A800M Instruct?

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 Q6_K_L (2.8 GB file, about 3.8 GB in use).
  • 6 GB VRAM: the best quant that fits is Q8_0 (3.6 GB file, about 4.6 GB in use).
  • 8 GB VRAM: the best quant that fits is F16 (6.8 GB file, about 7.8 GB in use).

Which Granite 3.0 3b A800M Instruct quant should you choose?

Q4_K_M is the usual starting point for Granite 3.0 3b A800M Instruct: it balances file size and quality well. Here it is a 2.1 GB file.

The Q8_0 file is 3.6 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 Q2_K at 1.3 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

Architecturegranitemoe
Parameters3.37B (approximate, mixture of experts)
Layers32
Trained context4,096 tokens
KV cache (fp16)64 KB per token

Source

Quantized bybartowski
FamilyGranite
Licenseapache-2.0
Downloads (HF)1,562
Last updated2024-10-21

Frequently asked questions about Granite 3.0 3b A800M Instruct

How much VRAM does Granite 3.0 3b A800M Instruct need?

About 3.1 GB at Q4_K_M with an 8K context window. That is the 2.1 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 Granite 3.0 3b A800M Instruct on an 8 GB GPU?

Yes, with F16: the 6.8 GB file needs about 7.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Granite 3.0 3b A800M Instruct GGUF?

It depends on the quant: Q2_K is 1.3 GB, Q2_K_L is 1.3 GB, IQ3_XS is 1.4 GB, Q3_K_S is 1.5 GB, IQ3_M is 1.5 GB, Q3_K_M is 1.6 GB and 17 more.

What context length does Granite 3.0 3b A800M Instruct support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 0.5 GB at 8K and 0.3 GB at 4K.

Where can I download Granite 3.0 3b A800M Instruct GGUF?

The files are on Hugging Face in the bartowski/granite-3.0-3b-a800m-instruct-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 Granite 3.0 3b A800M Instruct 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, a Q4_K_M quant uses about 3.1 GB VRAM at 8k context, fitting within 8 GB.

Does the model support vision tasks?

No, the model has no vision capability.

What license applies to the model?

The model is released under the Apache 2.0 license.

Other GGUFs of granite-3.0-3b-a800m-instruct

Similar models

Browse all →

More from bartowski

All by bartowski →

More Granite models

All Granite →

Other tiny models

All tiny models →