PY

PygWin 4x7B

PygwinGGUF by mradermacherUpdated 2024-05-06
Download on Hugging Face
Parameters19.7B
Context4K
Layers32
Architecturellama
Smallest quant7.4 GB
Quick answer

PygWin 4x7B needs about 16.9 GB of memory at Q4_K_M (a 12.1 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 14 quants, from 7.4 GB (Q2_K) to 21.1 GB (Q8_0).

PygWin 4x7B – a 19.7B parameter Mixture of Experts LLaMA model

PygWin 4x7B is a 19.7B parameter LLaMA‑based MoE model. It combines four 7B experts from Pygmalion and Xwin. The model supports a 4096 token context window and is released under the llama2 license.

Good at

  • Large context reasoning
  • Multi‑expert inference
  • High‑capacity language modeling

Limits

  • No vision support
  • Requires 8k context for memory estimates
  • License restricts commercial use

Best for

  • Research prototyping
  • High‑capacity text generation
  • Fine‑tuning on large corpora

Who it’s for. Researchers and developers needing a large MoE model for text tasks. The model is suitable for environments with sufficient GPU memory and a need for high‑capacity inference.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K7.4 GB 10.0 GB
Too big Get file →
IQ3_XS8.2 GB 10.8 GB
Too big Get file →
IQ3_S8.7 GB 11.3 GB
Too big Get file →
Q3_K_S8.7 GB 11.3 GB
Too big Get file →
IQ3_M8.9 GB 11.6 GB
Too big Get file →
Q3_K_M9.6 GB 12.3 GB
Too big Get file →
Q3_K_L10.4 GB 13.1 GB
Too big Get file →
IQ4_XS10.8 GB 13.4 GB
Too big Get file →
Q4_K_S11.4 GB 14.0 GB
Too big Get file →
Q4_K_M12.1 GB 14.7 GB
Too big Get file →
Q5_K_S13.7 GB 16.4 GB
Too big Get file →
Q5_K_M14.2 GB 16.8 GB
Too big Get file →
Q6_K16.3 GB 19.0 GB
Too big Get file →
Q8_021.1 GB 23.7 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (7.39 GB, 12.2 GB memory at 8k context). For 12 GB VRAM, Q4_K_S (11.38 GB, 16.2 GB memory) is a good balance. For 24 GB VRAM, Q6_K (16.34 GB, 21.1 GB memory) or Q8_0 (21.08 GB, 25.9 GB memory) provide higher quality.

What GPU do you need for PygWin 4x7B?

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: the best quant that fits is IQ4_XS (10.8 GB file, about 15.6 GB in use).
  • 24 GB VRAM: the best quant that fits is Q6_K (16.3 GB file, about 21.1 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (21.1 GB file, about 25.9 GB in use).

Which PygWin 4x7B quant should you choose?

Q4_K_M is the usual starting point for PygWin 4x7B: it balances file size and quality well. Here it is a 12.1 GB file.

The Q8_0 file is 21.1 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 Q2_K at 7.4 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

Architecturellama
Parameters19.7B (approximate, mixture of experts)
Layers32
Trained context4,096 tokens
KV cache (fp16)512 KB per token

Source

Quantized bymradermacher
FamilyPygwin
Licensellama2
Downloads (HF)2,156
Last updated2024-05-06

Frequently asked questions about PygWin 4x7B

How much VRAM does PygWin 4x7B need?

About 16.9 GB at Q4_K_M with an 8K context window. That is the 12.1 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 PygWin 4x7B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 12.2 GB, so you would need a larger GPU or CPU offloading.

What is the file size of PygWin 4x7B GGUF?

It depends on the quant: Q2_K is 7.4 GB, IQ3_XS is 8.2 GB, IQ3_S is 8.7 GB, Q3_K_S is 8.7 GB, IQ3_M is 8.9 GB, Q3_K_M is 9.6 GB and 8 more.

What context length does PygWin 4x7B support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 4.3 GB at 8K and 2.1 GB at 4K.

Where can I download PygWin 4x7B GGUF?

The files are on Hugging Face in the mradermacher/PygWin-4x7B-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 PygWin 4x7B use?

Hugging Face lists the license as llama2. 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 PygWin 4x7B on 8 GB VRAM?

Use the Q2_K quant, which is 7.39 GB and requires about 12.2 GB memory at 8k context.

Does the model support image inputs?

No, the model has no vision capability.

What license applies to this model?

The model is released under the llama2 license.

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Pygwin models

All Pygwin →

Other medium models

All medium models →