MP

Mpt 30b Instruct

MptGGUF by mradermacherUpdated 2024-09-19
Download on Hugging Face
Parameters30B
Context8K
Layers48
Architecturempt
Smallest quant11.3 GB
Quick answer

Mpt 30b Instruct needs about 30.9 GB of memory at Q4_K_M (a 19.1 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 14 quants, from 11.3 GB (Q2_K) to 31.8 GB (Q8_0).

Mpt 30b Instruct – 30 billion‑parameter instruction‑tuned model

Mpt 30b Instruct is a 30 billion‑parameter instruction‑tuned model based on the Mpt architecture. It supports a context length of 8192 tokens and is released under the Apache‑2.0 license. The model is available in a variety of GGUF quantizations.

Good at

  • Instruction following
  • Large context handling
  • High‑parameter capacity

Limits

  • No vision support
  • Requires large VRAM for higher‑quality quants
  • No explicit language list

Best for

  • Chatbot development
  • Content generation
  • Research prototyping

Who it’s for. Developers and researchers who need a large‑scale instruction‑tuned model for text generation tasks. The model is suitable for environments with sufficient GPU memory to run the desired quantization.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K11.3 GB 23.0 GB
Too big Get file →
IQ3_XS12.7 GB 24.4 GB
Too big Get file →
IQ3_S13.0 GB 24.8 GB
Too big Get file →
Q3_K_S13.0 GB 24.8 GB
Too big Get file →
IQ3_M14.5 GB 26.3 GB
Too big Get file →
Q3_K_M15.7 GB 27.5 GB
Too big Get file →
IQ4_XS16.2 GB 28.0 GB
Too big Get file →
Q4_K_S17.1 GB 28.9 GB
Too big Get file →
Q3_K_L17.2 GB 29.0 GB
Too big Get file →
Q4_K_M19.1 GB 30.9 GB
Too big Get file →
Q5_K_S20.6 GB 32.4 GB
Too big Get file →
Q5_K_M22.3 GB 34.1 GB
Too big Get file →
Q6_K24.6 GB 36.4 GB
Too big Get file →
Q8_031.8 GB 43.6 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (file size 11.26 GB, estimated memory 23 GB) or IQ3_XS (12.67 GB, 24.4 GB). For 16 GB VRAM, Q4_K_S (17.1 GB, 28.9 GB) or Q5_K_S (20.65 GB, 32.4 GB) are suitable. Higher‑quality quants such as Q6_K or Q8_0 require more than 36 GB VRAM.

What GPU do you need for Mpt 30b 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: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: the best quant that fits is Q2_K (11.3 GB file, about 23.0 GB in use).
  • 32 GB VRAM: the best quant that fits is Q4_K_M (19.1 GB file, about 30.9 GB in use).
  • 48 GB VRAM: the best quant that fits is Q8_0 (31.8 GB file, about 43.6 GB in use).

Which Mpt 30b Instruct quant should you choose?

Q4_K_M is the usual starting point for Mpt 30b Instruct: it balances file size and quality well. Here it is a 19.1 GB file.

The Q8_0 file is 31.8 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 11.3 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

Architecturempt
Parameters30B
Layers48
Trained context8,192 tokens
KV cache (fp16)1344 KB per token

Source

Quantized bymradermacher
FamilyMpt
Licenseapache-2.0
Downloads (HF)2,177
Last updated2024-09-19

Frequently asked questions about Mpt 30b Instruct

How much VRAM does Mpt 30b Instruct need?

About 30.9 GB at Q4_K_M with an 8K context window. That is the 19.1 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 Mpt 30b Instruct on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 23.0 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Mpt 30b Instruct GGUF?

It depends on the quant: Q2_K is 11.3 GB, IQ3_XS is 12.7 GB, IQ3_S is 13.0 GB, Q3_K_S is 13.0 GB, IQ3_M is 14.5 GB, Q3_K_M is 15.7 GB and 8 more.

What context length does Mpt 30b Instruct support?

The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 11.3 GB at 8K and 11.3 GB at 8K.

Where can I download Mpt 30b Instruct GGUF?

The files are on Hugging Face in the mradermacher/mpt-30b-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 Mpt 30b 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 Mpt 30b Instruct on 8 GB VRAM?

Use the Q2_K or IQ3_XS quantizations; they estimate 23–24.4 GB memory at 8k context, which is above 8 GB but can be run with CPU fallback or smaller batch sizes.

What license applies to this model?

The model is released under the Apache‑2.0 license.

Does this model support vision tasks?

No, the model does not have vision capabilities.

Other GGUFs of mpt-30b-instruct

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Mpt models

All Mpt →

Other medium models

All medium models →