Instructions to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF",
	filename="zamba2-2.7b-instruct-v2-q4_0.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0
# Run inference directly in the terminal:
llama-cli -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0
# Run inference directly in the terminal:
llama-cli -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0
# Run inference directly in the terminal:
./llama-cli -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0

Use Docker

docker model run hf.co/EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0

LM Studio
Jan

vLLM

How to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0

Ollama
How to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with Ollama:
```
ollama run hf.co/EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0
```

Unsloth Studio new

How to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF to start chatting

Docker Model Runner
How to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with Docker Model Runner:
```
docker model run hf.co/EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0
```

Lemonade

How to use EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF:Q4_0

Run and chat with the model

lemonade run user.Zamba2-2.7B-Instruct-v2-GGUF-Q4_0

List all available models

lemonade list

Zamba2 2.7B Instruct v2 - GGUF

GGUF conversions of Zyphra/Zamba2-2.7B-instruct-v2 for use with llama.cpp.

Architecture

Zamba2 is a hybrid Mamba-2 + shared Transformer architecture by Zyphra.

2.7B parameters, 54 layers (9 hybrid attention blocks, 2 shared transformers cycling even/odd)
Hidden size: 2560, attention hidden: 5120
SSM: d_state=64, d_conv=4, ngroups=1
No RoPE (use_mem_rope=false) — attention without positional encoding
Requires llama.cpp with Zamba2 support (PR #21412)

Available Quantizations

Quant	Size	BPW	PPL (WikiText-2)	Prompt tok/s	Gen tok/s	Hardware
Q4_0	2.1 GB	4.56	17.53	280.9	26.6	RTX 4090
Q8_0	3.9 GB	8.51	16.94	278.3	26.0	RTX 4090

Perplexity Comparison (WikiText-2, n_ctx=512)

Config	Description	PPL
A	Q4_0 weights + F16 KV	17.53
D	Q8_0 weights + F16 KV	16.94
E	F32 weights + F16 KV	~16.8 (pending)
B/C	Q4/Q8 KV cache	Requires FA (head_dim=160)

Note on KV cache quantization: The 2.7B model has head_dim=160, which is not a multiple of 64. This prevents the Hadamard rotation used by llama.cpp's standard KV quantization. Quantized KV cache requires Flash Attention, which is not yet wired for Zamba2.

Bug Fix (v2 release — 2026-04-04)

These GGUFs replace an earlier release that had a critical bug. The 2.7B model has use_mem_rope=false in its HuggingFace config, meaning it does not use rotary position embeddings in attention. The original converter applied RoPE unconditionally, corrupting all attention computations. This was the ONLY Zamba2 size affected (1.2B and 7B both have use_mem_rope=true).

Before fix: PPL = 37.90 (Q4_0), 37.69 (Q8_0), 37.68 (F32)
After fix: PPL = 17.53 (Q4_0), 16.94 (Q8_0)
Now correctly ordered: 1.2B (22.5) > 2.7B (17.5) > 7B (13.8)

Sample Output (Q8_0)

Q: What is the capital of France?

A: The capital of France is Paris.

Usage

# Build llama.cpp with Zamba2 support
git clone https://github.com/echo313unfolding/llama.cpp -b zamba2-support
cd llama.cpp && cmake -B build -DGGML_CUDA=ON && cmake --build build -j

# Run
./build/bin/llama-cli -m zamba2-2.7b-instruct-v2-q8_0.gguf \
  -p "<|im_start|>user\nWhat is quantum computing?<|im_end|>\n<|im_start|>assistant\n" \
  -n 256 -ngl 999 -e --no-conversation

Conversion Details

Converted from HF safetensors using a custom Zamba2-to-GGUF converter:

Mamba-2 SSM layers: A_log to A conversion, conv1d squeeze, dt projection
2 shared transformer blocks with per-layer LoRA unfolding (W_eff = W_shared + B @ A)
Per-layer n_head_kv array (0 for Mamba layers, 32 for attention layers)
rope_dimension_count=0 for use_mem_rope=false models (disables RoPE in attention)
BPE tokenizer (v2 format)
F32 master, quantized with llama-quantize

Credits

Architecture: Zyphra (Zamba2)
GGUF Conversion: Echo Labs
llama.cpp Zamba2 support: PR #21412

Downloads last month: 47

GGUF

Model size

4B params

Architecture

zamba2

Hardware compatibility

4-bit

8-bit