FI

v6 Finch 14B HF

FinchGGUF by bartowskiUpdated 2024-09-14
Download on Hugging Face
Parameters14.1B
Context1024K
Layers61
Architecturerwkv6
Smallest quant5.5 GB
Quick answer

v6 Finch 14B HF needs about 10.3 GB of memory at Q4_K_M (a 8.8 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 24 quants, from 5.5 GB (IQ2_M) to 28.6 GB (F16).

v6 Finch 14B HF – a 14.1B‑parameter RWKV model in GGUF format

v6 Finch 14B HF is a 14.1B‑parameter RWKV model released under the apache-2.0 license. It supports a context length of 1,048,576 tokens and is available in multiple GGUF quantizations. The model is designed for efficient local inference with llama.cpp.

Good at

  • Long‑context text generation
  • Low‑latency inference on GPUs
  • Fine‑tuning with small datasets

Limits

  • No vision capabilities
  • Requires 8‑k context for memory estimates
  • Large VRAM needed for higher‑precision quants

Best for

  • Chatbots needing long context
  • Research prototyping on local GPUs
  • Educational demos of RWKV architecture

Who it’s for. Developers and researchers who need a large‑parameter RWKV model for local inference. The model is suitable for those who prefer open‑source licensing and can manage the required VRAM or RAM resources.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_M5.5 GB 6.6 GB
Fits Get file →
Q2_K5.6 GB 6.8 GB
Fits Get file →
Q2_K_L5.8 GB 7.1 GB
Fits Get file →
IQ3_M7.0 GB 8.3 GB
Too big Get file →
IQ3_XS7.0 GB 8.3 GB
Too big Get file →
Q3_K_L7.0 GB 8.3 GB
Too big Get file →
Q3_K_M7.0 GB 8.3 GB
Too big Get file →
Q3_K_S7.0 GB 8.3 GB
Too big Get file →
Q3_K_XL7.2 GB 8.6 GB
Too big Get file →
IQ4_XS8.3 GB 9.8 GB
Too big Get file →
Q4_08.8 GB 10.3 GB
Too big Get file →
Q4_08.8 GB 10.3 GB
Too big Get file →
Q4_08.8 GB 10.3 GB
Too big Get file →
Q4_08.8 GB 10.3 GB
Too big Get file →
Q4_K_M8.8 GB 10.3 GB
Too big Get file →
Q4_K_S8.8 GB 10.3 GB
Too big Get file →
Q4_K_L9.0 GB 10.5 GB
Too big Get file →
Q5_K_M10.5 GB 12.2 GB
Too big Get file →
Q5_K_S10.5 GB 12.2 GB
Too big Get file →
Q5_K_L10.6 GB 12.4 GB
Too big Get file →
Q6_K12.3 GB 14.2 GB
Too big Get file →
Q6_K_L12.4 GB 14.4 GB
Too big Get file →
Q8_015.6 GB 18.0 GB
Too big Get file →
F1628.6 GB 32.5 GB
Too big Get file →

For 8‑GB VRAM, use Q4_K_S (file size 8.77 GB, ~10.3 GB memory). For 12‑GB VRAM, Q5_K_S (10.46 GB, ~12.2 GB memory) is a good balance. If you have 16‑GB VRAM, Q6_K (12.27 GB, ~14.2 GB memory) offers higher precision.

What GPU do you need for v6 Finch 14B HF?

These figures add the file size, an fp16 KV cache at an 8K context window 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: the best quant that fits is Q2_K_L (5.8 GB file, about 7.1 GB in use).
  • 12 GB VRAM: the best quant that fits is Q4_K_L (9.0 GB file, about 10.5 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K_L (12.4 GB file, about 14.4 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (15.6 GB file, about 18.0 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (15.6 GB file, about 18.0 GB in use).
  • 48 GB VRAM: the best quant that fits is F16 (28.6 GB file, about 32.5 GB in use).

Which v6 Finch 14B HF quant should you choose?

Q4_K_M is the usual starting point for v6 Finch 14B HF: it balances file size and quality well. Here it is a 8.8 GB file.

The Q8_0 file is 15.6 GB, about 1.8 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 IQ2_M at 5.5 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

Architecturerwkv6
Parameters14.1B
Layers61
Trained context1,048,576 tokens

Source

Quantized bybartowski
FamilyFinch
Licenseapache-2.0
Downloads (HF)1,096
Last updated2024-09-14

Frequently asked questions about v6 Finch 14B HF

How much VRAM does v6 Finch 14B HF need?

About 10.3 GB at Q4_K_M with an 8K context window. That is the 8.8 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 v6 Finch 14B HF on an 8 GB GPU?

Yes, with Q2_K_L: the 5.8 GB file needs about 7.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of v6 Finch 14B HF GGUF?

It depends on the quant: IQ2_M is 5.5 GB, Q2_K is 5.6 GB, Q2_K_L is 5.8 GB, IQ3_M is 7.0 GB, IQ3_XS is 7.0 GB, Q3_K_L is 7.0 GB and 18 more.

What context length does v6 Finch 14B HF support?

The GGUF header lists a trained context length of 1,048,576 tokens (1024K).

Where can I download v6 Finch 14B HF GGUF?

The files are on Hugging Face in the bartowski/v6-Finch-14B-HF-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 v6 Finch 14B HF 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 v6 Finch 14B HF on 8 GB VRAM?

Yes, with a Q4_K_S quant you can fit the model on 8 GB VRAM, though memory usage is approximate and may vary with batch size.

Does the model support image inputs?

No, v6 Finch 14B HF does not have vision capabilities.

What license applies to this model?

The model is released under the apache-2.0 license.

Other GGUFs of v6-Finch-14B-HF

Similar models

Browse all →

More from bartowski

All by bartowski →

More Finch models

All Finch →

Other medium models

All medium models →