How much VRAM does Chocolatine-3B-Instruct-DPO-Revised need?

Q4_K_M is exactly 2,393,231,840 bytes (2.23 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

How large is Chocolatine-3B-Instruct-DPO-Revised's KV cache?

12.00 GiB at 32K context with an f16 cache, computed per layer from the model's architecture. Quantizing the cache to q8_0 roughly halves it.

jpacifico · text

Chocolatine-3B-Instruct-DPO-Revised

jpacifico/Chocolatine-3B-Instruct-DPO-Revised

Chocolatine-3B-Instruct-DPO-Revised at Q4_K_M is exactly 2,393,231,840 bytes (2.23 GiB / 2.39 GB) — an effective 5.011 bits per weight, not the nominal 4. Its KV cache at 32K is 12.00 GiB.

From the file· summed from 1 file(s)From the file· KV per layer

Parameters

3.8B

Architecture

phi3

32 layers

Context

4,096

native (config.json)

License

mit

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q4_K_M	2.23 GiB	2,393,231,840	5.011	—	jpacifico

KV cache by context

computed per layer

Context	KV cache (f16)●	Flat formula	Overstated by	Full / windowed / recurrent
4,096	1.50 GiB	1.50 GiB	—	32 / 0 / 0
8,192	3.00 GiB	3.00 GiB	—	32 / 0 / 0
16,384	6.00 GiB	6.00 GiB	—	32 / 0 / 0
32,768	12.00 GiB	12.00 GiB	—	32 / 0 / 0
65,536	24.00 GiB	24.00 GiB	—	32 / 0 / 0
131,072	48.00 GiB	48.00 GiB	—	32 / 0 / 0

Compare with

same modality, comparable size

vs gemma-4-E4B-it vs Qwen3-4B vs Qwen3-8B vs HyperCLOVAX-SEED-Text-Instruct-1.5B vs Llama-3.2-1B-Instruct vs llama-3-youko-8b All alternatives →Will it fit on my card? →

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 2.00 GiB. The real file is 2.23 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json

Layers

Attention heads

KV heads

Head dim

Hidden size

3072

Vocab

32,064

Sliding window

2047

SWA period

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does Chocolatine-3B-Instruct-DPO-Revised need?: Q4_K_M is exactly 2,393,231,840 bytes (2.23 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is Chocolatine-3B-Instruct-DPO-Revised's KV cache?: 12.00 GiB at 32K context with an f16 cache, computed per layer. Quantizing the cache to q8_0 roughly halves it, which is often the difference between a context length fitting and not.
Which quantization of Chocolatine-3B-Instruct-DPO-Revised should I use?: Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.