KE

KernelGen LM 14B

KernelgenGGUF by mradermacherUpdated 2026-05-19
Download on Hugging Face
Parameters14.8B
Context40K
Layers40
Architectureqwen3
Smallest quant0.0 GB
Quick answer

KernelGen LM 14B needs about 10.8 GB of memory at Q4_K_M (a 9.0 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 25 quants, from 0.0 GB to 12.1 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 5.4 GB.

KernelGen LM 14B – domain‑specific NPU kernel generation model

KernelGen LM 14B is a 14.8 billion parameter model based on Qwen3. It is trained on the Ascend‑CoT dataset and fine‑tuned with reinforcement learning for low‑level NPU kernel generation. The model supports coding, reasoning, and tool usage.

Good at

  • Low‑level NPU kernel generation
  • Chain‑of‑Thought reasoning
  • Tool‑aware code synthesis

Limits

  • No vision support
  • Requires 8 k context for memory estimates
  • Apache‑2.0 license limits commercial use without attribution

Best for

  • Generating AscendC kernels
  • Optimizing NPU code for Huawei Ascend
  • Evaluating kernel performance on real hardware

Who it’s for. Developers and researchers working on Huawei Ascend hardware who need a specialized LLM for kernel generation and optimization. The model is suitable for those who can run GGUF files with llama.cpp or similar tools.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.0 GB 1.8 GB
Fits Get file →
IQ1_S3.6 GB 5.4 GB
Fits Get file →
IQ1_M3.8 GB 5.7 GB
Fits Get file →
IQ2_XXS4.3 GB 6.1 GB
Fits Get file →
IQ2_XS4.7 GB 6.5 GB
Fits Get file →
IQ2_S5.0 GB 6.8 GB
Fits Get file →
IQ2_M5.3 GB 7.2 GB
Fits Get file →
Q2_K_S5.4 GB 7.2 GB
Tight Get file →
Q2_K5.8 GB 7.6 GB
Tight Get file →
IQ3_XXS5.9 GB 7.8 GB
Tight Get file →
IQ3_XS6.4 GB 8.2 GB
Too big Get file →
Q3_K_S6.7 GB 8.5 GB
Too big Get file →
IQ3_S6.7 GB 8.5 GB
Too big Get file →
IQ3_M6.9 GB 8.7 GB
Too big Get file →
Q3_K_M7.3 GB 9.2 GB
Too big Get file →
Q3_K_L7.9 GB 9.7 GB
Too big Get file →
IQ4_XS8.1 GB 10.0 GB
Too big Get file →
IQ4_NL8.5 GB 10.4 GB
Too big Get file →
Q4_08.5 GB 10.4 GB
Too big Get file →
Q4_K_S8.6 GB 10.4 GB
Too big Get file →
Q4_K_M9.0 GB 10.8 GB
Too big Get file →
Q4_19.4 GB 11.2 GB
Too big Get file →
Q5_K_S10.3 GB 12.1 GB
Too big Get file →
Q5_K_M10.5 GB 12.4 GB
Too big Get file →
Q6_K12.1 GB 14.0 GB
Too big Get file →

For 8 GB VRAM, use Q4_K_M (9 GB file, ~10.8 GB memory). For 16 GB VRAM, Q5_K_M (10.51 GB file, ~12.4 GB memory) offers higher quality. If memory is limited, IQ1_S (3.58 GB file, ~5.4 GB memory) is a lightweight option.

What GPU do you need for KernelGen LM 14B?

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: the best quant that fits is the main file (0.0 GB file, about 1.8 GB in use).
  • 6 GB VRAM: the best quant that fits is IQ1_M (3.8 GB file, about 5.7 GB in use).
  • 8 GB VRAM: the best quant that fits is IQ3_XXS (5.9 GB file, about 7.8 GB in use).
  • 12 GB VRAM: the best quant that fits is Q4_1 (9.4 GB file, about 11.2 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K (12.1 GB file, about 14.0 GB in use).

Which KernelGen LM 14B quant should you choose?

Q4_K_M is the usual starting point for KernelGen LM 14B: it balances file size and quality well. Here it is a 9.0 GB file.

If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen3
Parameters14.8B
Layers40
Trained context40,960 tokens
KV cache (fp16)160 KB per token

Source

Quantized bymradermacher
FamilyKernelgen
Licenseapache-2.0
Downloads (HF)687
Last updated2026-05-19

Frequently asked questions about KernelGen LM 14B

How much VRAM does KernelGen LM 14B need?

About 10.8 GB at Q4_K_M with an 8K context window. That is the 9.0 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 KernelGen LM 14B on an 8 GB GPU?

Yes, with IQ3_XXS: the 5.9 GB file needs about 7.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of KernelGen LM 14B GGUF?

It depends on the quant: file is 0.0 GB, IQ1_S is 3.6 GB, IQ1_M is 3.8 GB, IQ2_XXS is 4.3 GB, IQ2_XS is 4.7 GB, IQ2_S is 5.0 GB and 19 more.

What context length does KernelGen LM 14B support?

The GGUF header lists a trained context length of 40,960 tokens (40K). The KV cache grows with the context you use: about 1.3 GB at 8K and 5.4 GB at 32K.

Where can I download KernelGen LM 14B GGUF?

The files are on Hugging Face in the mradermacher/KernelGen-LM-14B-i1-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 KernelGen LM 14B 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 KernelGen LM 14B on 8 GB VRAM?

Yes, the Q4_K_M quant uses about 10.8 GB memory at 8 k context, which fits within 8 GB VRAM with some overhead.

Does the model support vision tasks?

No, the model has no vision capabilities.

What license applies to this model?

The model is released under the Apache‑2.0 license.

Other GGUFs of KernelGen-LM-14B

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Kernelgen models

All Kernelgen →

Other medium models

All medium models →