ZE

LORAs

ZephyrGGUF by kukidevalmlUpdated 2026-06-10
Download on Hugging Face
Parameters4B
Context40K
Layers36
Architectureqwen3
Smallest quant8.1 GB
Quick answer

LORAs needs about 9.8 GB of memory at F16 (a 8.1 GB file) with an 8K context window, so it fits on a 12 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 4.8 GB.

LORAs: 4B parameter Zephyr fine‑tuned LoRA model

LORAs is a 4B parameter model based on Zephyr‑7B‑β, fine‑tuned with LoRA techniques. It is released under the Apache‑2.0 license and supports a 40960 token context window. The model is available in a single F16 quantized GGUF file.

Good at

  • Chat‑style interactions
  • Fine‑tuned LoRA performance
  • Large context handling

Limits

  • No vision support
  • Requires 9.8 GB VRAM at 8 k context
  • Single quant available

Best for

  • Chatbot prototypes
  • Fine‑tuning experiments
  • Context‑heavy dialogue systems

Who it’s for. Developers and researchers who need a lightweight Zephyr‑based model for chat applications or LoRA fine‑tuning. The model is suitable for projects that can accommodate the 9.8 GB VRAM requirement and do not need vision capabilities.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
F168.1 GB 9.8 GB
Too big Get file →

The only available quant is F16, with a file size of 8.05 GB and an estimated VRAM usage of 9.8 GB at 8 k context. This is the recommended choice for systems with at least 10 GB of VRAM. No other quantizations are provided.

What GPU do you need for LORAs?

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 F16 (8.1 GB file, about 9.8 GB in use).

Which LORAs quant should you choose?

F16 (8.1 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen3
Parameters4B
Layers36
Trained context40,960 tokens
KV cache (fp16)144 KB per token

Source

Quantized bykukidevalml
FamilyZephyr
Licenseapache-2.0
Downloads (HF)1,318
Last updated2026-06-10

Frequently asked questions about LORAs

How much VRAM does LORAs need?

About 9.8 GB at F16 with an 8K context window. That is the 8.1 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 LORAs on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 9.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of LORAs GGUF?

It depends on the quant: F16 is 8.1 GB.

What context length does LORAs support?

The GGUF header lists a trained context length of 40,960 tokens (40K). The KV cache grows with the context you use: about 1.2 GB at 8K and 4.8 GB at 32K.

Where can I download LORAs GGUF?

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

What license does LORAs 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 LORAs on 8 GB VRAM?

No, the F16 quant requires approximately 9.8 GB VRAM at 8 k context, exceeding an 8 GB limit.

Does LORAs support vision tasks?

No, the model does not include vision capabilities.

Other GGUFs of zephyr-7b-beta

Similar models

Browse all →

More from kukidevalml

All by kukidevalml →

More Zephyr models

All Zephyr →

Other small models

All small models →