Downloads · 30 days
68
3% of all-time downloads
gemma-4-12B-it-assistant-GGUF
IHaveNoClueAndIMustPost/gemma-4-12B-it-assistant-GGUF
gemma-4-12B-it-assistant-GGUF is a machine learning model from IHaveNoClueAndIMustPost. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
FP16 test draft model conversion for google/gemma-4-12B-it
All-time downloads
2.2K
Public
Repo size
1.7 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.gguf862 MB · 100%
At a glance
- Access
- Public
- Created
- Jun 4, 2026
- Updated
- Jun 4, 2026
- SHA
- 0e80ddb0
Base models
Model card
From the Hugging Face model README
FP16 test draft model conversion for google/gemma-4-12B-it
Surprisingly "--spec-draft-n-max 1" looks like the winner but it could be due to a whole slew of factors including a poor draft quantization by me.
| Task | Baseline (tok/s) | Draft N=1 (tok/s) | Draft N=2 (tok/s) | Draft N=3 (tok/s) | Draft N=4 (tok/s) |
|---|---|---|---|---|---|
| code_python | 49.3 | 51.5 | 46.5 | 45.6 | 50.1 |
| code_cpp | 49.2 | 60.3 | 55.8 | 54.6 | 50.9 |
| explain_concept | 49.2 | 57.1 | 54.4 | 49.9 | 46.8 |
| summarize | 48.7 | 60.8 | 61.3 | 62.8 | 55.9 |
| qa_factual | 48.6 | 59.5 | 55.9 | 56.8 | 55.5 |
| translation | 48.8 | 59.0 | 53.9 | 59.2 | 49.5 |
| creative_short | 48.0 | 52.3 | 46.0 | 46.1 | 41.4 |
| stepwise_math | 48.4 | 57.4 | 55.4 | 60.6 | 49.8 |
| long_code_review | 47.7 | 52.8 | 53.9 | 52.0 | 44.4 |
| Aggregate Avg. | ~48.6 | ~56.5 | ~53.7 | ~54.3 | ~49.5 |
- Created
- Jun 4, 2026
- Updated
- Jun 4, 2026