Potato-CODER-12GB-Fast needs about 9.5 GB of memory at its main file (a 8.3 GB file) with an 8K context window, so it fits on a 12 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 2.7 GB.
Potato-CODER-12GB-Fast: a 35B ternary Qwen model optimized for fast decoding on NVIDIA GPUs
Potato-CODER-12GB-Fast is a 35B-A3B Qwen model with 90% ternary weights. It runs entirely on the GPU with no CPU weight offloading and supports vision encoding. The model is tuned for interactive coding tasks and fast inference.
Good at
- Fast decoding on NVIDIA GPUs
- Interactive coding with vision
- Large context support up to 262,144 tokens
Limits
- Requires NVIDIA GPU with 570+ driver
- No CPU weight offloading
- No support for non‑NVIDIA hardware
Best for
- Rapid code generation
- Vision‑enabled coding assistants
- Interactive debugging sessions
Who it’s for. Developers and researchers who need a quick, GPU‑only inference solution for coding tasks with vision support. The model is ideal for interactive work where speed is critical.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 8.3 GB | 9.5 GB | Too big | Get file → |
The only available GGUF file is the ternary quant with a file size of 8.31 GB. It requires approximately 9.5 GB of VRAM at 8 k context. For 8 GB VRAM, the model will not fit; consider a smaller quant or a different model.
What GPU do you need for Potato-CODER-12GB-Fast?
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: the best quant that fits is the main file (8.3 GB file, about 9.5 GB in use).
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Potato-CODER-12GB-Fast
How much VRAM does Potato-CODER-12GB-Fast need?
About 9.5 GB at its main file with an 8K context window. That is the 8.3 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 Potato-CODER-12GB-Fast on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 9.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Potato-CODER-12GB-Fast GGUF?
It depends on the quant: file is 8.3 GB.
What context length does Potato-CODER-12GB-Fast support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.7 GB at 8K and 2.7 GB at 32K.
Where can I download Potato-CODER-12GB-Fast GGUF?
The files are on Hugging Face in the Slomin/Potato-CODER-12GB-Fast repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Potato-CODER-12GB-Fast 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 Potato-CODER-12GB-Fast on an 8 GB VRAM GPU?
No. The model needs about 9.5 GB of VRAM at 8 k context, so an 8 GB GPU is insufficient.
Does the model support vision tasks?
Yes. It includes a vision projector that runs on the GPU.
What driver version is required?
An NVIDIA driver 570+ is required for Linux; 551+ for Windows.