Question 1

How much VRAM does HiLS-Attention-7B need?

Accepted Answer

HiLS-Attention-7B (7.3B) requires approximately 16.349908447265623 GB of VRAM at full precision (FP16), 8.2 GB at 8-bit quantization (Q8), and 4.1 GB at 4-bit quantization (Q4). The actual VRAM usage may be slightly higher due to KV cache and runtime overhead.

Question 2

What GPUs can run HiLS-Attention-7B?

Accepted Answer

We have 10 compatible GPUs indexed for HiLS-Attention-7B, including A40, A100 40GB, A10G, RTX A6000, A10 and 5 more. Compatibility is determined by whether the GPU's VRAM can fit the model's memory footprint, potentially using quantization to reduce memory usage.

Question 3

What is the cheapest GPU to run HiLS-Attention-7B?

Accepted Answer

The cheapest GPU to run HiLS-Attention-7B in the cloud is the A40 at $0.44/hr. Other affordable options include A100 40GB ($0.93/hr) and A10G ($1.01/hr). Prices reflect the lowest on-demand rate across cloud providers.

Question 4

Can HiLS-Attention-7B run on consumer hardware?

Accepted Answer

Yes, HiLS-Attention-7B can run on consumer GPUs like the RTX A6000 (48 GB VRAM) at full precision (FP16).

Question 5

What quantization options are available for HiLS-Attention-7B?

Accepted Answer

HiLS-Attention-7B supports the following quantization levels: FP16 (full precision): ~16.349908447265623 GB, Q8 (8-bit): ~8.2 GB, Q6 (6-bit): ~6.1 GB, Q4 (4-bit): ~4.1 GB. Lower quantization reduces VRAM usage at the cost of some quality loss. Q8 offers a good balance between quality and memory savings, while Q4 provides maximum compression.

GPU	VRAM	Best Quant	From	Rent
A40	48 GB	FP16	$0.44/hr	Rent
A100 40GB	40 GB	FP16	$0.93/hr	Rent
A10G	24 GB	FP16	$1.01/hr
RTX A6000	48 GB	FP16	$1.09/hr
A10	24 GB	FP16	$1.29/hr
A100 80GB PCIe	80 GB	FP16	$1.39/hr	Rent
A100 80GB	80 GB	FP16	$1.39/hr	Rent
GH200	96 GB	FP16	$2.29/hr
H100 MEGA	80 GB	FP16	$2.89/hr	Rent
H100 SXM	80 GB	FP16	$2.89/hr	Rent

HiLS-Attention-7B

VRAM Requirements

Compatible GPUs

HiLS-Attention-7B FAQ

VRAM requirements for HiLS-Attention-7B at different quantization levels
Quantization	VRAM Required
FP16	16.3 GB
Q8	8.2 GB
Q6	6.1 GB
Q4	4.1 GB