MA

madlad400 10b mt

MadladGGUF by thirteenbitUpdated 2024-07-06
Download on Hugging Face
Parameters10.7B
Context512
Layers32
Architecturet5
Smallest quant5.2 GB
Quick answer

madlad400 10b mt needs about 11.5 GB of memory at Q4_K_M (a 6.7 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 5 quants, from 5.2 GB (Q3_K_M) to 11.4 GB (Q8_0).

Madlad400 10b mt – a 10.7B‑parameter multilingual T5 model in GGUF format

Madlad400 10b mt is a multilingual machine translation model based on the T5 architecture. It contains 10.7 billion parameters and was trained on 250 billion tokens covering over 450 languages. The model is released under the Apache‑2.0 license.

Good at

  • Machine translation across 400+ languages
  • Multilingual NLP tasks
  • Research‑grade baseline model

Limits

  • No vision support
  • Requires 8 k context for memory estimates
  • Not evaluated for production use

Best for

  • Translating documents between low‑resource languages
  • Multilingual text generation experiments
  • Benchmarking multilingual LLMs

Who it’s for. Researchers and developers needing a large multilingual translation model. The model is suitable for academic studies and prototype development, but is not intended for production deployments.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q3_K_M5.2 GB 6.0 GB
Fits Get file →
Q4_K_M6.7 GB 7.5 GB
Tight Get file →
Q5_K_M7.7 GB 8.5 GB
Too big Get file →
Q6_K8.8 GB 9.6 GB
Too big Get file →
Q8_011.4 GB 12.2 GB
Too big Get file →

For 8 GB VRAM, use Q3_K_M (file size 5.22 GB, estimated 10 GB memory). For 12 GB VRAM, Q4_K_M (6.69 GB, 11.5 GB memory) is a good balance. For 16 GB VRAM, Q5_K_M (7.71 GB, 12.5 GB memory) or Q6_K (8.79 GB, 13.6 GB memory) can be used. The estimates are approximate.

What GPU do you need for madlad400 10b mt?

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: the best quant that fits is Q4_K_M (6.7 GB file, about 11.5 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K (8.8 GB file, about 13.6 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (11.4 GB file, about 16.2 GB in use).

Which madlad400 10b mt quant should you choose?

Q4_K_M is the usual starting point for madlad400 10b mt: it balances file size and quality well. Here it is a 6.7 GB file.

The Q8_0 file is 11.4 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 Q3_K_M at 5.2 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturet5
Parameters10.7B
Layers32
Trained context512 tokens
KV cache (fp16)512 KB per token

Source

Quantized bythirteenbit
FamilyMadlad
Licenseapache-2.0
Downloads (HF)1,496
Last updated2024-07-06

Frequently asked questions about madlad400 10b mt

How much VRAM does madlad400 10b mt need?

About 11.5 GB at Q4_K_M with an 8K context window. That is the 6.7 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 madlad400 10b mt on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 10.0 GB, so you would need a larger GPU or CPU offloading.

What is the file size of madlad400 10b mt GGUF?

It depends on the quant: Q3_K_M is 5.2 GB, Q4_K_M is 6.7 GB, Q5_K_M is 7.7 GB, Q6_K is 8.8 GB, Q8_0 is 11.4 GB.

What context length does madlad400 10b mt support?

The GGUF header lists a trained context length of 512 tokens (512). The KV cache grows with the context you use: about 4.3 GB at 8K and 0.3 GB at 512.

Where can I download madlad400 10b mt GGUF?

The files are on Hugging Face in the thirteenbit/madlad400-10b-mt-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 madlad400 10b mt 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 madlad400 10b mt on 8 GB VRAM?

Yes, the Q3_K_M quantization requires about 10 GB of VRAM at 8 k context, which is close to 8 GB but may need GPU offloading.

Does this model support vision tasks?

No, madlad400 10b mt is a language model and does not include vision capabilities.

What license applies to this model?

The model is released under the Apache‑2.0 license, allowing free use and modification.

Other GGUFs of madlad400-10b-mt

Similar models

Browse all →

More from thirteenbit

All by thirteenbit →

More Madlad models

All Madlad →

Other medium models

All medium models →