RA

Ralphseek v4 Flash Mtp Native Hf v2

RalphseekGGUF by ji-farthingUpdated 2026-08-13
Download on Hugging Face
Parameters122M
Context1024K
Layers5
Architecturedeepseek4
Smallest quant0.0 GB
Quick answer

Ralphseek v4 Flash Mtp Native Hf v2 needs about 0.7 GB of memory at IQ4_NL (a 0.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 6 quants, from 0.0 GB (IQ4_NL) to 0.5 GB (F32). Longer context needs more: at 32K tokens the KV cache alone adds about 0.3 GB.

Ralphseek v4 Flash Mtp Native Hf v2 is a 122M-parameter Ralphseek model in GGUF format, quantized by ji-farthing. Its trained context length is 1024K tokens. The repository has 6 quant files, from 0.0 GB (IQ4_NL) to 0.5 GB (F32).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ4_NL0.0 GB 0.6 GB
Fits Get file →
Q8_00.1 GB 0.6 GB
Fits Get file →
IQ4_NL0.1 GB 0.7 GB
Fits Get file →
Q8_00.1 GB 0.7 GB
Fits Get file →
F320.2 GB 0.8 GB
Fits Get file →
F320.5 GB 1.1 GB
Fits Get file →

What GPU do you need for Ralphseek v4 Flash Mtp Native Hf v2?

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 F32 (0.5 GB file, about 1.1 GB in use).

Which Ralphseek v4 Flash Mtp Native Hf v2 quant should you choose?

IQ4_NL (0.1 GB) is the middle option in this repository. The usual sweet spot between file size and quality, and the quant most people start with.

If memory is tight, the smallest file is IQ4_NL at 0.0 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

Architecturedeepseek4
Parameters122M (approximate, mixture of experts)
Layers5
Trained context1,048,576 tokens
KV cache (fp16)10 KB per token

Source

Quantized byji-farthing
FamilyRalphseek
Licensemit
Downloads (HF)186
Last updated2026-08-13

Frequently asked questions about Ralphseek v4 Flash Mtp Native Hf v2

How much VRAM does Ralphseek v4 Flash Mtp Native Hf v2 need?

About 0.7 GB at IQ4_NL with an 8K context window. That is the 0.1 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 Ralphseek v4 Flash Mtp Native Hf v2 on an 8 GB GPU?

Yes, with F32: the 0.5 GB file needs about 1.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Ralphseek v4 Flash Mtp Native Hf v2 GGUF?

It depends on the quant: IQ4_NL is 0.0 GB, Q8_0 is 0.1 GB, IQ4_NL is 0.1 GB, Q8_0 is 0.1 GB, F32 is 0.2 GB, F32 is 0.5 GB.

What context length does Ralphseek v4 Flash Mtp Native Hf v2 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 0.1 GB at 8K and 0.3 GB at 32K.

Where can I download Ralphseek v4 Flash Mtp Native Hf v2 GGUF?

The files are on Hugging Face in the ji-farthing/RalphSeek-V4-Flash-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 Ralphseek v4 Flash Mtp Native Hf v2 use?

Hugging Face lists the license as mit. 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 ji-farthing

All by ji-farthing →

More Ralphseek models

All Ralphseek →

Other tiny models

All tiny models →