TW

Twilight-Miqu-146B

TwilightGGUF by mradermacherUpdated 2024-08-02
Download on Hugging Face
Parameters146B
Context32K
Layers170
Architecturellama
Smallest quant30.6 GB
Quick answer

Twilight-Miqu-146B needs about 44.8 GB of memory at IQ2_XXS (a 38.6 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 6 quants, from 30.6 GB (IQ1_S) to 49.0 GB (IQ2_M).

Large story‑writing model with 146B parameters

Twilight‑Miqu‑146B is a 146‑billion‑parameter llama model fine‑tuned for creative story writing. It supports up to 32,764 context tokens and is available in several GGUF quantizations.

Good at

  • Creative story narration
  • Coherent dialogue generation
  • Iterative plot development

Limits

  • Context limited to 32,764 tokens
  • No vision capabilities
  • License type: other

Best for

  • Writing fiction
  • Collaborative storytelling
  • Creative writing assistance

Who it’s for. Developers and writers who need an AI assistant for crafting detailed, coherent stories. The model is suitable for projects that require extensive narrative generation and iterative editing.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S30.6 GB 36.8 GB
Too big Get file →
IQ1_M33.6 GB 39.8 GB
Too big Get file →
IQ2_XXS38.6 GB 44.8 GB
Too big Get file →
IQ2_XS42.9 GB 49.1 GB
Too big Get file →
IQ2_S45.0 GB 51.2 GB
Too big Get file →
IQ2_M49.0 GB 55.2 GB
Too big Get file →

For 8 GB VRAM, use IQ1_S (30.58 GB file, ~36.8 GB memory at 8 k context). For 12 GB VRAM, IQ1_M (33.57 GB file, ~39.8 GB memory). For 16 GB VRAM, IQ2_XXS (38.56 GB file, ~44.8 GB memory). Memory estimates are approximate.

What GPU do you need for Twilight-Miqu-146B?

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: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: the best quant that fits is IQ2_XXS (38.6 GB file, about 44.8 GB in use).

Which Twilight-Miqu-146B quant should you choose?

IQ2_XXS (38.6 GB) is the middle option in this repository. 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.

If memory is tight, the smallest file is IQ1_S at 30.6 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

Architecturellama
Parameters146B
Layers170
Trained context32,764 tokens
KV cache (fp16)680 KB per token

Source

Quantized bymradermacher
FamilyTwilight
Licenseother
Downloads (HF)636
Last updated2024-08-02

Frequently asked questions about Twilight-Miqu-146B

How much VRAM does Twilight-Miqu-146B need?

About 44.8 GB at IQ2_XXS with an 8K context window. That is the 38.6 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 Twilight-Miqu-146B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 36.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Twilight-Miqu-146B GGUF?

It depends on the quant: IQ1_S is 30.6 GB, IQ1_M is 33.6 GB, IQ2_XXS is 38.6 GB, IQ2_XS is 42.9 GB, IQ2_S is 45.0 GB, IQ2_M is 49.0 GB.

What context length does Twilight-Miqu-146B support?

The GGUF header lists a trained context length of 32,764 tokens (32K). The KV cache grows with the context you use: about 5.7 GB at 8K and 22.8 GB at 32K.

Where can I download Twilight-Miqu-146B GGUF?

The files are on Hugging Face in the mradermacher/Twilight-Miqu-146B-i1-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 Twilight-Miqu-146B use?

Hugging Face lists the license as other. 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 Twilight‑Miqu‑146B on 8 GB VRAM?

Use the IQ1_S quant; it requires about 36.8 GB memory at 8 k context, so 8 GB VRAM is insufficient.

Does the model support image input?

No, the model has no vision capabilities.

What is the maximum context length?

The model supports up to 32,764 tokens.

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Twilight models

All Twilight →

Other large models

All large models →