Wan2.2-Animate-14B · QuantStack

Wan2.2-Animate-14B Q4_0

This file is exactly 10,402,699,072 bytes — 9.69 GiB / 10.40 GB at an effective 4.817 bits per weight. The nominal rate for Q4_0 is lower; mixed-precision tensors make the real figure higher, always.

From the file· 1 file(s), summedFrom the file· 1441 tensors parsed

Get it

9.69 GiB · 1 file
llama.cpp
llama-cli -hf QuantStack/Wan2.2-Animate-14B-GGUF:Q4_0

Downloads and runs in one step, resolving the quantization by name.

Hugging Face CLI
hf download QuantStack/Wan2.2-Animate-14B-GGUF Wan2.2-Animate-14B-Q4_0.gguf
Direct download

Straight from the Hugging Face CDN — we host nothing and earn nothing from this. Verify what you received against the exact byte count above; a size mismatch is the usual cause of a file that will not load.

Repository →
Size
9.69 GiB
10.40 GB
Effective bpw
4.817
from real bytes ÷ params
Tensors
1441
Header
0.09 MB
GGUF metadata

What this quantization actually contains

per-tensor types, parsed from the GGUF header
F32
929
Q4_0
464
Q4_1
40
BF16
8

A quantization label names a mixture, not a uniform precision. Attention and output tensors are routinely kept at higher precision than the label implies, which is exactly why the effective bits-per-weight above exceeds the nominal rate.

Fit by accelerator

AcceleratorMemory○Total @4K◐Total @32K◐Fits◐tok/s @4K◐
Arc A310 4GB4 GB10.52 GiB10.52 GiBno—
Arc A350M 4GB4 GB10.52 GiB10.52 GiBno—
Arc A370M 4GB4 GB10.52 GiB10.52 GiBno—
Arc A530M 4GB4 GB10.52 GiB10.52 GiBno—
Arc Pro A30M 4GB4 GB10.52 GiB10.52 GiBno—
Radeon Pro W64004 GB10.62 GiB10.62 GiBno—
Radeon RX 64004 GB10.62 GiB10.62 GiBno—
Radeon RX 6500 XT4 GB10.62 GiB10.62 GiBno—
RTX A4004 GB10.72 GiB10.72 GiBno—
Arc A380 6GB6 GB10.52 GiB10.52 GiBno—
Arc Pro A40 6GB6 GB10.52 GiB10.52 GiBno—
Arc Pro A50 6GB6 GB10.52 GiB10.52 GiBno—
GeForce RTX 20606 GB10.52 GiB10.52 GiBno—
GeForce RTX 30506 GB10.52 GiB10.52 GiBno—
GeForce RTX 3060 OEM6 GB10.52 GiB10.52 GiBno—
RTX A20006 GB10.72 GiB10.72 GiBno—
Apple M18 GB10.27 GiB10.27 GiBno—
Apple M28 GB10.27 GiB10.27 GiBno—
Apple M38 GB10.27 GiB10.27 GiBno—
Apple M48 GB10.27 GiB10.27 GiBno—
Arc A530M 8GB8 GB10.52 GiB10.52 GiBno—
Arc A550M 8GB8 GB10.52 GiB10.52 GiBno—
Arc A570M 8GB8 GB10.52 GiB10.52 GiBno—
Arc A580 8GB8 GB10.52 GiB10.52 GiBno—
○Spec sheet◐Predictedwhat these mean

Memory at context

ContextWeights●KV (f16)●KV (q8_0)●Working buffer◐Total (f16)◐
●From the file◐Predictedwhat these mean

Quantizing the KV cache to q8_0 is a roughly 2× lever on the dominant term at long context, and it is the single most useful setting most local users never touch. Totals here exclude the allocator reserve your driver takes, which is hardware-specific — the per-accelerator table above includes it.