SW

Swallow 70b Instruct v0.1

SwallowGGUF by RichardErkhovUpdated 2024-07-28
Download on Hugging Face
Parameters69B
Context4K
Layers80
Architecturellama
Smallest quant25.6 GB
Quick answer

Swallow 70b Instruct v0.1 needs about 44.7 GB of memory at Q4_K_M (a 41.5 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 22 quants, from 25.6 GB (Q2_K) to 73.5 GB (Q8_0).

Swallow 70b Instruct v0.1 – 70 billion‑parameter instruction‑tuned LLaMA model

Swallow 70b Instruct v0.1 is a 70 billion‑parameter LLaMA‑based model tuned for instruction following. It supports Japanese and English text. The model is released under the llama2 license.

Good at

  • Instruction following
  • Japanese and English text
  • Large context handling

Limits

  • Requires large VRAM for full precision
  • No vision support
  • License restricts commercial use

Best for

  • Chatbot development
  • Content generation
  • Research prototyping

Who it’s for. Developers and researchers needing a high‑capacity instruction model for Japanese and English. Suitable for environments with sufficient GPU memory and compliance with the llama2 license.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K25.6 GB 27.4 GB
Too big Get file →
IQ3_XS28.4 GB 30.3 GB
Too big Get file →
IQ3_S30.0 GB 31.9 GB
Too big Get file →
Q3_K_S30.0 GB 31.9 GB
Too big Get file →
IQ3_M31.1 GB 32.9 GB
Too big Get file →
Q3_K33.4 GB 35.2 GB
Too big Get file →
Q3_K_M33.4 GB 35.2 GB
Too big Get file →
Q3_K_L36.3 GB 38.1 GB
Too big Get file →
IQ4_XS37.3 GB 39.2 GB
Too big Get file →
Q4_039.0 GB 40.8 GB
Too big Get file →
IQ4_NL39.4 GB 41.2 GB
Too big Get file →
Q4_K_S39.4 GB 41.2 GB
Too big Get file →
Q4_K41.5 GB (2 parts) 43.4 GB
Too big Get file →
Q4_K_M41.5 GB (2 parts) 43.4 GB
Too big Get file →
Q4_143.3 GB (2 parts) 45.1 GB
Too big Get file →
Q5_047.6 GB (2 parts) 49.4 GB
Too big Get file →
Q5_K_S47.6 GB (2 parts) 49.4 GB
Too big Get file →
Q6_K47.6 GB (2 parts) 49.5 GB
Too big Get file →
Q5_K48.9 GB (2 parts) 50.7 GB
Too big Get file →
Q5_K_M48.9 GB (2 parts) 50.7 GB
Too big Get file →
Q5_151.9 GB (2 parts) 53.7 GB
Too big Get file →
Q8_073.5 GB (2 parts) 75.3 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (25.57 GB file, 28.8 GB memory). For 16 GB VRAM, Q4_0 (39 GB file, 42.2 GB memory) or Q4_K (41.55 GB file, 44.7 GB memory). For 24 GB VRAM, Q5_0 (47.6 GB file, 50.8 GB memory) or Q5_K (48.89 GB file, 52.1 GB memory).

What GPU do you need for Swallow 70b Instruct v0.1?

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: the best quant that fits is IQ3_XS (28.4 GB file, about 31.6 GB in use).
  • 48 GB VRAM: the best quant that fits is Q4_1 (43.3 GB file, about 46.5 GB in use).

Which Swallow 70b Instruct v0.1 quant should you choose?

Q4_K_M is the usual starting point for Swallow 70b Instruct v0.1: it balances file size and quality well. Here it is a 41.5 GB file.

The Q8_0 file is 73.5 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 Q2_K at 25.6 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
Parameters69B
Layers80
Trained context4,096 tokens
KV cache (fp16)320 KB per token

Source

Quantized byRichardErkhov
FamilySwallow
Licensellama2
Downloads (HF)2,761
Last updated2024-07-28

Frequently asked questions about Swallow 70b Instruct v0.1

How much VRAM does Swallow 70b Instruct v0.1 need?

About 44.7 GB at Q4_K_M with an 8K context window. That is the 41.5 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 Swallow 70b Instruct v0.1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 28.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Swallow 70b Instruct v0.1 GGUF?

It depends on the quant: Q2_K is 25.6 GB, IQ3_XS is 28.4 GB, IQ3_S is 30.0 GB, Q3_K_S is 30.0 GB, IQ3_M is 31.1 GB, Q3_K is 33.4 GB and 16 more.

What context length does Swallow 70b Instruct v0.1 support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 2.7 GB at 8K and 1.3 GB at 4K.

Where can I download Swallow 70b Instruct v0.1 GGUF?

The files are on Hugging Face in the RichardErkhov/tokyotech-llm_-_Swallow-70b-instruct-v0.1-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 Swallow 70b Instruct v0.1 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 Swallow 70b Instruct v0.1 on 8 GB VRAM?

The Q2_K quantization requires about 28.8 GB VRAM, so 8 GB is insufficient.

Does the model support vision tasks?

No, the model only handles text.

What license applies to this model?

The model is released under the llama2 license.

Similar models

Browse all →

More from RichardErkhov

All by RichardErkhov →

More Swallow models

All Swallow →

Other large models

All large models →