JI

JingyaoLi MoTCoder 15B v1.0

JingyaoliGGUF by RichardErkhovUpdated 2024-10-31
Download on Hugging Face
Parameters15.5B
Context8K
Layers40
Architecturestarcoder
Smallest quant2.8 GB
Quick answer

JingyaoLi MoTCoder 15B v1.0 needs about 10.6 GB of memory at Q4_K_M (a 10.0 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 19 quants, from 2.8 GB (IQ4_XS) to 16.6 GB (Q8_0).

JingyaoLi MoTCoder 15B v1.0 is a 15.5B-parameter Jingyaoli model in GGUF format, quantized by RichardErkhov. Its trained context length is 8K tokens. The repository has 19 quant files, from 2.8 GB (IQ4_XS) to 16.6 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ4_XS2.8 GB 3.5 GB
Fits Get file →
Q3_K_L4.5 GB 5.2 GB
Fits Get file →
Q2_K6.2 GB 6.9 GB
Fits Get file →
Q3_K_S7.0 GB 7.6 GB
Tight Get file →
Q3_K_M7.4 GB 8.0 GB
Too big Get file →
Q3_K8.2 GB 8.9 GB
Too big Get file →
Q4_09.0 GB 9.7 GB
Too big Get file →
IQ4_NL9.1 GB 9.8 GB
Too big Get file →
Q4_K_S9.1 GB 9.8 GB
Too big Get file →
Q4_19.9 GB 10.6 GB
Too big Get file →
Q4_K10.0 GB 10.6 GB
Too big Get file →
Q4_K_M10.0 GB 10.6 GB
Too big Get file →
Q5_010.9 GB 11.6 GB
Too big Get file →
Q5_K_S10.9 GB 11.6 GB
Too big Get file →
Q5_K11.5 GB 12.2 GB
Too big Get file →
Q5_K_M11.5 GB 12.2 GB
Too big Get file →
Q5_111.8 GB 12.5 GB
Too big Get file →
Q6_K12.9 GB 13.6 GB
Too big Get file →
Q8_016.6 GB 17.3 GB
Too big Get file →

What GPU do you need for JingyaoLi MoTCoder 15B v1.0?

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 IQ4_XS (2.8 GB file, about 3.5 GB in use).
  • 6 GB VRAM: the best quant that fits is Q3_K_L (4.5 GB file, about 5.2 GB in use).
  • 8 GB VRAM: the best quant that fits is Q3_K_M (7.4 GB file, about 8.0 GB in use).
  • 12 GB VRAM: the best quant that fits is Q5_0 (10.9 GB file, about 11.6 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K (12.9 GB file, about 13.6 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (16.6 GB file, about 17.3 GB in use).

Which JingyaoLi MoTCoder 15B v1.0 quant should you choose?

Q4_K_M is the usual starting point for JingyaoLi MoTCoder 15B v1.0: it balances file size and quality well. Here it is a 10.0 GB file.

The Q8_0 file is 16.6 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 IQ4_XS at 2.8 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturestarcoder
Parameters15.5B
Layers40
Trained context8,192 tokens
KV cache (fp16)20 KB per token

Source

Quantized byRichardErkhov
FamilyJingyaoli
Downloads (HF)123
Last updated2024-10-31

Frequently asked questions about JingyaoLi MoTCoder 15B v1.0

How much VRAM does JingyaoLi MoTCoder 15B v1.0 need?

About 10.6 GB at Q4_K_M with an 8K context window. That is the 10.0 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 JingyaoLi MoTCoder 15B v1.0 on an 8 GB GPU?

Yes, with Q3_K_M: the 7.4 GB file needs about 8.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of JingyaoLi MoTCoder 15B v1.0 GGUF?

It depends on the quant: IQ4_XS is 2.8 GB, Q3_K_L is 4.5 GB, Q2_K is 6.2 GB, Q3_K_S is 7.0 GB, Q3_K_M is 7.4 GB, Q3_K is 8.2 GB and 13 more.

What context length does JingyaoLi MoTCoder 15B v1.0 support?

The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 0.2 GB at 8K and 0.2 GB at 8K.

Where can I download JingyaoLi MoTCoder 15B v1.0 GGUF?

The files are on Hugging Face in the RichardErkhov/JingyaoLi_-_MoTCoder-15B-v1.0-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 JingyaoLi MoTCoder 15B v1.0 use?

Hugging Face lists the license as bigscience-openrail-m. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Similar models

Browse all →

More from RichardErkhov

All by RichardErkhov →

Other medium models

All medium models →