QW

Qwen2.5 95B Instruct i1

QwenGGUF by mradermacherUpdated 2024-10-23
Download on Hugging Face
Parameters94.6B
Context32K
Layers105
Architectureqwen2
Smallest quant29.4 GB
Quick answer

Qwen2.5 95B Instruct i1 needs about 42.6 GB of memory at Q2_K (a 38.6 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 13 quants, from 29.4 GB (IQ1_S) to 48.9 GB (Q3_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 14.1 GB.

Qwen2.5 95B Instruct i1 is a 94.6B-parameter Qwen model in GGUF format, quantized by mradermacher. Its trained context length is 32K tokens. The repository has 13 quant files, from 29.4 GB (IQ1_S) to 48.9 GB (Q3_K_M).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S29.4 GB 33.4 GB
Too big Get file →
IQ1_M30.8 GB 34.8 GB
Too big Get file →
IQ2_XXS33.1 GB 37.1 GB
Too big Get file →
IQ2_XS35.1 GB 39.1 GB
Too big Get file →
IQ2_S36.2 GB 40.3 GB
Too big Get file →
IQ2_M38.1 GB 42.1 GB
Too big Get file →
Q2_K38.6 GB 42.6 GB
Too big Get file →
IQ3_XXS41.2 GB 45.3 GB
Too big Get file →
IQ3_XS42.5 GB 46.5 GB
Too big Get file →
Q3_K_S44.5 GB 48.6 GB
Too big Get file →
IQ3_S44.7 GB 48.7 GB
Too big Get file →
IQ3_M46.0 GB 50.0 GB
Too big Get file →
Q3_K_M48.9 GB 52.9 GB
Too big Get file →

What GPU do you need for Qwen2.5 95B Instruct i1?

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 IQ3_XS (42.5 GB file, about 46.5 GB in use).

Which Qwen2.5 95B Instruct i1 quant should you choose?

Q2_K (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 29.4 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

Architectureqwen2
Parameters94.6B
Layers105
Trained context32,768 tokens
KV cache (fp16)420 KB per token

Source

Quantized bymradermacher
FamilyQwen
Licenseother
Downloads (HF)84
Last updated2024-10-23

Frequently asked questions about Qwen2.5 95B Instruct i1

How much VRAM does Qwen2.5 95B Instruct i1 need?

About 42.6 GB at Q2_K 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 Qwen2.5 95B Instruct i1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 33.4 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Qwen2.5 95B Instruct i1 GGUF?

It depends on the quant: IQ1_S is 29.4 GB, IQ1_M is 30.8 GB, IQ2_XXS is 33.1 GB, IQ2_XS is 35.1 GB, IQ2_S is 36.2 GB, IQ2_M is 38.1 GB and 7 more.

What context length does Qwen2.5 95B Instruct i1 support?

The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 3.5 GB at 8K and 14.1 GB at 32K.

Where can I download Qwen2.5 95B Instruct i1 GGUF?

The files are on Hugging Face in the mradermacher/Qwen2.5-95B-Instruct-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 Qwen2.5 95B Instruct i1 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.

Other GGUFs of Qwen2.5-95B-Instruct

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Qwen models

All Qwen →

Other large models

All large models →