Run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive locally
HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive is a mid-size language model with 7.52 billion parameters, built on the gemma4 architecture. It is released under the gemma license and has been downloaded 573,160 times.
To run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive locally at a 4,096-token context, its quantized versions need between 1.91 GB (F16, lowest quality) and 8.56 GB (Q8_K_P, highest quality) of memory, weights plus KV cache and a system margin included.
For most users the best balance is Q6_K_P, needing about 6.81 GB. That means HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive fits entirely in the VRAM of a 6 GB GPU or larger, running fully on the GPU.
All quantizations
| Quant. | Bits | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| F16 | 1.05 | Very low | 0.92 GB | 0.19 GB | 1.91 GB | 433.7 t/s | Fits in VRAM |
| Q2_K_P | 4.72 | Good | 4.13 GB | 0.19 GB | 5.12 GB | 96.9 t/s | Fits in VRAM |
| IQ3_M | 5.02 | Very good | 4.39 GB | 0.19 GB | 5.38 GB | 91.1 t/s | Fits in VRAM |
| Q3_K_M | 5.16 | Very good | 4.52 GB | 0.19 GB | 5.5 GB | 88.5 t/s | Fits in VRAM |
| Q3_K_P | 5.2 | Very good | 4.55 GB | 0.19 GB | 5.54 GB | 87.9 t/s | Fits in VRAM |
| IQ4_XS | 5.4 | Very good | 4.72 GB | 0.19 GB | 5.71 GB | 84.7 t/s | Fits in VRAM |
| Q4_K_M | 5.68 | Very good | 4.97 GB | 0.19 GB | 5.96 GB | 80.5 t/s | Fits in VRAM |
| Q4_K_P | 5.71 | Very good | 5.0 GB | 0.19 GB | 5.99 GB | 80.0 t/s | Fits in VRAM |
| Q5_K_M | 6.13 | Very good | 5.37 GB | 0.19 GB | 6.35 GB | 74.5 t/s | Fits in VRAM |
| Q5_K_P | 6.19 | Very good | 5.41 GB | 0.19 GB | 6.4 GB | 73.9 t/s | Fits in VRAM |
| Q6_K_P | 6.65 | Excellent | 5.82 GB | 0.19 GB | 6.81 GB | 68.7 t/s | Fits in VRAM |
| Q8_K_P | 8.65 | Excellent | 7.57 GB | 0.19 GB | 8.56 GB | 6.6 t/s | Offload |
KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.
Frequently asked questions
How much VRAM do you need to run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive?
You need about 5.99 GB of VRAM to run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive entirely on the GPU using the Q4_K_P quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.
Can I run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive on an 8 GB GPU?
Yes. With 8 GB of VRAM you can run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive fully on the GPU using Q6_K_P (about 6.81 GB).
Can I run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive on a 16 GB GPU?
Yes. With 16 GB of VRAM you can run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive fully on the GPU using Q8_K_P (about 8.56 GB).
Can I run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive on a 24 GB GPU?
Yes. With 24 GB of VRAM you can run HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive fully on the GPU using Q8_K_P (about 8.56 GB).
What is the best quantization for HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive?
If memory allows, higher bits-per-weight means better quality. A common sweet spot is a Q4_K_M or Q5_K_M quantization, which keeps most of the quality while roughly halving the memory versus 8-bit. Pick the highest quantization that still fits in your VRAM.