MI

miniG

MinigGGUF by bartowskiUpdated 2024-08-27
Download on Hugging Face
Parameters9.5B
Context1024K
Layers40
Architecturechatglm
Smallest quant4.0 GB
Quick answer

miniG needs about 7.5 GB of memory at Q4_K_M (a 6.3 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 24 quants, from 4.0 GB (IQ2_M) to 19.0 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 2.7 GB.

miniG – 9.5B chatglm model in GGUF format

miniG is a 9.5B parameter chatglm model available in multiple GGUF quantizations. It supports up to 1,048,576 context tokens. The model is provided by bartowski and can be used with llama.cpp or LM Studio.

Good at

  • Chat‑style dialogue
  • Large context handling
  • Low‑memory inference

Limits

  • No vision support
  • License not specified
  • Requires 8‑k context for memory estimates

Best for

  • Local chatbot deployment
  • Fine‑tuning on small datasets
  • Testing quantization effects

Who it’s for. Developers and researchers who need a compact chatglm model for local inference. The model is suitable for experimentation with different quantization schemes.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_M4.0 GB 5.1 GB
Fits Get file →
Q2_K4.0 GB 5.2 GB
Fits Get file →
IQ3_XS4.5 GB 5.6 GB
Fits Get file →
Q3_K_S4.6 GB 5.8 GB
Fits Get file →
Q2_K_L4.6 GB 5.8 GB
Fits Get file →
IQ3_M4.9 GB 6.0 GB
Fits Get file →
Q3_K_M5.1 GB 6.3 GB
Fits Get file →
IQ4_XS5.3 GB 6.5 GB
Fits Get file →
Q3_K_L5.3 GB 6.5 GB
Fits Get file →
Q4_05.5 GB 6.7 GB
Fits Get file →
Q4_05.5 GB 6.7 GB
Fits Get file →
Q4_05.5 GB 6.7 GB
Fits Get file →
Q4_05.5 GB 6.7 GB
Fits Get file →
Q4_K_S5.8 GB 7.0 GB
Fits Get file →
Q3_K_XL5.9 GB 7.0 GB
Fits Get file →
Q4_K_M6.3 GB 7.5 GB
Tight Get file →
Q5_K_S6.8 GB 7.9 GB
Tight Get file →
Q4_K_L6.8 GB 7.9 GB
Tight Get file →
Q5_K_M7.2 GB 8.4 GB
Too big Get file →
Q5_K_L7.6 GB 8.8 GB
Too big Get file →
Q6_K8.3 GB 9.5 GB
Too big Get file →
Q6_K_L8.6 GB 9.8 GB
Too big Get file →
Q8_010.1 GB 11.3 GB
Too big Get file →
F1619.0 GB 20.1 GB
Too big Get file →

For 8 GB VRAM, use IQ2_M (file size 3.96 GB, 5.1 GB memory). For 12 GB VRAM, Q4_K_M (6.31 GB, 7.5 GB memory) is a good balance. For 16 GB VRAM, Q5_K_L (7.6 GB, 8.8 GB memory) offers higher quality.

What GPU do you need for miniG?

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 IQ3_M (4.9 GB file, about 6.0 GB in use).
  • 8 GB VRAM: the best quant that fits is Q4_K_L (6.8 GB file, about 7.9 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (10.1 GB file, about 11.3 GB in use).
  • 16 GB VRAM: the best quant that fits is Q8_0 (10.1 GB file, about 11.3 GB in use).
  • 24 GB VRAM: the best quant that fits is F16 (19.0 GB file, about 20.1 GB in use).

Which miniG quant should you choose?

Q4_K_M is the usual starting point for miniG: it balances file size and quality well. Here it is a 6.3 GB file.

The Q8_0 file is 10.1 GB, about 1.6 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 IQ2_M at 4.0 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

Architecturechatglm
Parameters9.5B
Layers40
Trained context1,048,576 tokens
KV cache (fp16)80 KB per token

Source

Quantized bybartowski
FamilyMinig
Downloads (HF)1,188
Last updated2024-08-27

Frequently asked questions about miniG

How much VRAM does miniG need?

About 7.5 GB at Q4_K_M with an 8K context window. That is the 6.3 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 miniG on an 8 GB GPU?

Yes, with Q4_K_L: the 6.8 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 miniG GGUF?

It depends on the quant: IQ2_M is 4.0 GB, Q2_K is 4.0 GB, IQ3_XS is 4.5 GB, Q3_K_S is 4.6 GB, Q2_K_L is 4.6 GB, IQ3_M is 4.9 GB and 18 more.

What context length does miniG support?

The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 0.7 GB at 8K and 2.7 GB at 32K.

Where can I download miniG GGUF?

The files are on Hugging Face in the bartowski/miniG-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Can I run miniG on 8 GB VRAM?

Yes, use IQ2_M or Q2_K which fit within 8 GB VRAM and require about 5.1–5.2 GB memory at 8k context.

Does miniG support image inputs?

No, miniG is a text‑only model with no vision capabilities.

What is the maximum context length?

miniG supports up to 1,048,576 tokens in its context window.

Similar models

Browse all →

More from bartowski

All by bartowski →

Other small models

All small models →