SakuraLLM
Sakura-GalTransl-7B-v3.7
- Parameters
- 7.7B
- Max context length
- 32,768
- Architecture
- qwen2
- License
- Non-commercial
- License
- cc-by-nc-sa-4.0
About Sakura-GalTransl-7B-v3.7
Sakura-GalTransl-7B-v3.7 is a model published in GGUF format. GGUF files are published by SakuraLLM/Sakura-GalTransl-7B-v3.7. It has about 7.7B parameters, the qwen2 architecture, a maximum context length of 32,768 tokens (values written in the GGUF).
There are 3 quantizations, with file sizes from 4.25 GB (IQ4_XS) to 6.34 GB (Q6_K) (actual sizes on Hugging Face).
Running IQ4_XS (4.25 GB) with an 8,192-token context needs an estimated 9.62 GB of memory, and an estimated 22.5 GB at 32,768 tokens. For 8,192 tokens, a GPU with 12 GB or more of VRAM (such as GeForce RTX 3080 Ti, GeForce RTX 5070), or a Mac with 12 GB or more of unified memory, is a rough guide (macOS caps how much memory the GPU can use, so leave some headroom).
According to SakuraLLM/Sakura-GalTransl-7B-v3.7, the license is Creative Commons Attribution-NonCommercial-ShareAlike 4.0 (CC BY-NC-SA 4.0), which does not allow commercial use (limited to non-commercial use such as research or personal use). Check the original license for the terms of use.
File size and estimated memory per quantization
- Up to 8 GB
- Up to 16 GB
- Up to 24 GB
- Up to 48 GB
- Up to 128 GB
- More
File sizes are the actual values on Hugging Face. Memory is estimated for each context length (KV cache f16). The formula is in How memory is calculated.
All memory figures are Estimated (KV cache f16). It should run on a setup with at least this much VRAM (or Mac unified memory). Click a value to see the devices and models that fit that amount (Find by GPU / Mac).
About the "quality guide" (how far a quantization drifts from the original)
For each quantization type, how far the output drifts from the original model, grouped into six levels from values llama.cpp measured on Llama-3-8B.
The value is mean KLD (the difference between the original and quantized model in next-token probabilities); smaller means closer to the original. Quantizations made with an imatrix (importance data) come out slightly smaller. Levels use the value without an imatrix (or with one, for types that only have that).
These values were not measured on this model. Larger models and MoE models tend to drift less with the same type. Distributor-specific mixed types (such as UD-Q4_K_XL) cannot be matched by type name alone, so they show "-".
| Quantization | Quality guide | KLD (no imatrix) | KLD (imatrix) | Size (Llama-3-8B) |
|---|---|---|---|---|
| Q8_0 | Nearly identical | 0.0014 | - | 7.96 GiB |
| Q6_K | Very small | 0.0055 | - | 6.14 GiB |
| Q5_K_M | Very small | 0.0108 | - | 5.33 GiB |
| Q5_K_S | Small | 0.0166 | - | 5.21 GiB |
| Q5_1 | Small | 0.0180 | - | 5.65 GiB |
| Q5_0 | Small | 0.0222 | - | 5.21 GiB |
| Q4_K_M | Small | 0.0313 | 0.0282 | 4.58 GiB |
| Q4_K_S | Medium | 0.0431 | 0.0320 | 4.37 GiB |
| IQ4_NL | Small | - | 0.0357 | 4.35 GiB |
| IQ4_XS | Small | - | 0.0363 | 4.14 GiB |
| Q4_1 | Medium | 0.0717 | - | 4.78 GiB |
| Q4_0 | Medium | 0.0719 | - | 4.34 GiB |
| Q3_K_L | Medium | 0.0872 | 0.0731 | 4.03 GiB |
| Q3_K_M | Medium | 0.1019 | 0.0844 | 3.74 GiB |
| IQ3_M | Medium | - | 0.1025 | 3.53 GiB |
| IQ3_S | Medium | - | 0.1113 | 3.42 GiB |
| IQ3_XS | Large | - | 0.1387 | 3.28 GiB |
| IQ3_XXS | Large | - | 0.1836 | 3.05 GiB |
| Q3_K_S | Large | 0.2282 | 0.1998 | 3.41 GiB |
| IQ2_M | Large | - | 0.3260 | 2.74 GiB |
| Q2_K | Very large | 0.4451 | 0.3322 | 2.96 GiB |
| Q2_K_S | Very large | - | 0.4034 | 2.96 GiB |
| IQ2_S | Very large | - | 0.4392 | 2.56 GiB |
| IQ2_XS | Very large | - | 0.5463 | 2.43 GiB |
| IQ2_XXS | Very large | - | 0.8120 | 2.24 GiB |
| IQ1_M | Very large | - | 1.3931 | 2.01 GiB |
| IQ1_S | Very large | - | 2.2546 | 1.88 GiB |
Source: llama.cpp (MIT), measured on Llama-3-8B against FP16, Wikitext-2, revision f364eb6f. Original table · Reading and choosing quantizations
Last checked: Nov 11, 2026
- Architecture
- qwen2
- Supported before llama.cpp b4409
- Parameters
- 7.7B
- Max context length
- 32,768
- License
- cc-by-nc-sa-4.0 Non-commercial
| Quantization | File size Measured | Quality guide | 4,096 | 8,192 | 16,384 | 32,768 max |
|---|---|---|---|---|---|---|
| IQ4_XS | 4.25 GB | Small | 7.47 GB | 9.62 GB | 13.9 GB | 22.5 GB |
| Q5_K_S Sakura-GalTransl-7B-v3-Q5_K_S | 5.32 GB | Small | 8.54 GB | 10.7 GB | 15.0 GB | 23.6 GB |
| Q5_K_S Sakura-GalTransl-7B-v3.5-Q5_K_S | 5.32 GB | Small | 8.54 GB | 10.7 GB | 15.0 GB | 23.6 GB |
| Q6_K GalTransl-7B-v1-Q6_K | 6.34 GB | Very small | 9.56 GB | 11.7 GB | 16.0 GB | 24.6 GB |
| Q6_K GalTransl-7B-v1.5-Q6_K | 6.34 GB | Very small | 9.56 GB | 11.7 GB | 16.0 GB | 24.6 GB |
| Q6_K GalTransl-7B-v2-Q6_K | 6.34 GB | Very small | 9.56 GB | 11.7 GB | 16.0 GB | 24.6 GB |
| Q6_K GalTransl-7B-v2.6-Q6_K | 6.34 GB | Very small | 9.56 GB | 11.7 GB | 16.0 GB | 24.6 GB |
Ollama ollama run hf.co/SakuraLLM/Sakura-GalTransl-7B-v3.7:IQ4_XS
llama.cpp llama-server -hf SakuraLLM/Sakura-GalTransl-7B-v3.7:IQ4_XS
Generation speed on common GPUs and Macs (estimated)
A rough guide to how fast IQ4_XS (4.25 GB) generates text. Each token reads the model weights once, so the upper limit is memory bandwidth ÷ file size. The guide is 60–80% of that, where most published measurements fall.
| Device | Memory bandwidth | Guide (tokens/s) Estimated | Upper limit |
|---|---|---|---|
| Radeon RX 7600 XT 16 GB | 288 GB/s | 41–54 | 68 |
| Radeon RX 6950 XT 16 GB | 576 GB/s | 81–108 | 136 |
| GeForce RTX 5090 32 GB | 1,792 GB/s | 253–337 | 422 |
| Apple M2 16 GB | 100 GB/s | 14–19 | 24 |
| Apple M2 Pro 16 GB | 200 GB/s | 28–38 | 47 |
| Apple M5 Ultra 96 GB | 1,200 GB/s | 169–226 | 282 |
Estimates only. Prompt processing speed (limited by compute) is not included. Longer contexts and offloading part of the model to the CPU make it slower. Macs are shown only when installed memory is at least 4/3 of what is needed. Devices without published bandwidth (some GeForce cards, etc.) are not shown.
License and commercial use
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 (CC BY-NC-SA 4.0) Non-commercial
You may use, modify and redistribute for non-commercial purposes only, with attribution. Adaptations must be shared under the same (or a compatible) license. Commercial use is not allowed (non-commercial means not primarily intended for or directed towards commercial advantage or monetary compensation).
The original model's license has not been fetched yet, so this is based on the GGUF repository's label. Original license text (checked Nov 1, 2026)
The summary states facts this site read from the original text and is not legal advice. Always check the original license and its latest terms before use. All licenses
How to run
Example commands using IQ4_XS. To use another quantization, replace the name with one from the table above. How to run local LLMs
Ollama
Run it by pointing directly at the GGUF on Hugging Face.
ollama run hf.co/SakuraLLM/Sakura-GalTransl-7B-v3.7:IQ4_XS
llama.cpp
Pass the repository and quantization name to llama-server (or llama-cli) with -hf.
llama-server -hf SakuraLLM/Sakura-GalTransl-7B-v3.7:IQ4_XS
LM Studio
Paste SakuraLLM/Sakura-GalTransl-7B-v3.7 into the search box, download it, and choose the quantization from the file list.
Models of a similar size (other developers)
-
9.2B
ornith-ai
Ornith-1.5-9B
- Q4_K_M size
- 5.78 GB
- Memory at 8K
- 7.17 GB
Measured Estimated VRAM 8 GB+
-
9B
ornith-ai
Ornith-1.0-9B
- Q4_K_M size
- 5.63 GB
- Memory at 8K
- 7.02 GB
Measured Estimated VRAM 8 GB+
-
9B
DavidAU
Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP
- Q4_K_M size
- 6.83 GB
- Memory at 8K
- 8.20 GB
Measured Estimated VRAM 8 GB+
-
OpenMOSS-Team
MOSS-Music-8B-Instruct
- F16 size
- 1.72 GB
- Memory at 8K
- 3.78 GB
Measured Estimated VRAM 4 GB+
-
8.2B
Qwen
Qwen3-VL-8B-Instruct
- Q4_K_M size
- 5.03 GB
- Memory at 8K
- 7.08 GB
Measured Estimated VRAM 8 GB+
-
9B
Qwen
Qwen3.5-9B
- Q4_K_M size
- 5.68 GB
- Memory at 8K
- 7.05 GB
Measured Estimated VRAM 8 GB+
Frequently asked questions
- How much memory do I need to run Sakura-GalTransl-7B-v3.7 (IQ4_XS)?
- The file size is 4.25 GB (measured). With a 8,192-token context, the estimated memory including the KV cache is 9.62 GB. The formula is published at https://localllm.az-arte.com/en/guide/memory-calculation.
- Can Sakura-GalTransl-7B-v3.7 be used commercially?
- The license is Creative Commons Attribution-NonCommercial-ShareAlike 4.0 (CC BY-NC-SA 4.0) (No commercial use).You may use, modify and redistribute for non-commercial purposes only, with attribution. Adaptations must be shared under the same (or a compatible) license. Commercial use is not allowed (non-commercial means not primarily intended for or directed towards commercial advantage or monetary compensation). Check the original license text before deciding.
- Does Sakura-GalTransl-7B-v3.7 run in 8 GB, 12 GB, 16 GB or 24 GB of VRAM?
- 8 GB: no quantization fits; 12 GB: up to Q6_K (est. 11.7 GB); 16 GB: up to Q6_K (est. 11.7 GB); 24 GB: up to Q6_K (est. 11.7 GB) (estimated at 8,192 tokens with an f16 KV cache, files from the most-downloaded source).
- What is the command to run Sakura-GalTransl-7B-v3.7 with Ollama?
- Run "ollama run hf.co/SakuraLLM/Sakura-GalTransl-7B-v3.7:IQ4_XS" to use the GGUF on Hugging Face directly. To use another quantization, replace the name at the end.