---
title: Z1T-0-mlx
canonical_url: "https://www.modelscope.cn/models/mlx-community/Z1T-0-mlx"
md_url: "https://www.modelscope.cn/models/mlx-community/Z1T-0-mlx.md"
repository: mlx-community/Z1T-0-mlx
last_updated: 2026-09-05
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - z1t
base_model:
  - Extropic-AI/Z1T-0
base_model_relation: finetune
parameters: 1.2B
tensor_type:
  - F32
  - I32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 14
stars: 0
tags:
  - mlx
  - text-generation
  - attention-free
---

# Z1T-0-mlx

> Z1T-0-mlx - mlx-community 在 ModelScope 开源的模型。MLX conversion of Extropic-AI/Z1T-0 — the first open-weight Z1T, an attention-free, DyT-normed, fixed-sparsity decoder LM designed for Extropic's Z1 probabilistic hardware (research post, code).

mlx-community/Z1T-0-mlx 是 ModelScope 魔搭社区上的 1.2B 参数text-generation模型，采用 apache-2.0 许可，基于 Extropic-AI/Z1T-0 构建。

- **Repository**: mlx-community/Z1T-0-mlx
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 1.2B
- **Base model**: Extropic-AI/Z1T-0
- **Tags**: mlx, text-generation, attention-free
- **Downloads**: 14
- **Stars**: 0
- **Last updated**: 2026-09-05

Source: https://www.modelscope.cn/models/mlx-community/Z1T-0-mlx

---

# Z1T-0 (MLX)

MLX conversion of [**Extropic-AI/Z1T-0**](https://huggingface.co/Extropic-AI/Z1T-0) — the
first open-weight Z1T, an *attention-free*, DyT-normed, fixed-sparsity decoder LM designed for
Extropic's Z1 probabilistic hardware ([research post](https://extropic.ai/writing/z1t),
[code](https://github.com/extropic-ai/sparse-transformers)).

The original is a JAX/Equinox checkpoint (`.eqx`) with a bespoke architecture (no softmax
attention: causal depthwise conv + causal cumulative pool; `tanh`-linear fixed fan-in-4 sparse
projections; Dynamic-Tanh as the only norm). This repo ports it to MLX and ships it as
safetensors + a GPT-2 tokenizer.

**Faithfulness:** greedy decoding is **token-for-token identical** to the JAX reference; logits
match to `max|Δ|≈1.4e-2` (float32 accumulation-order noise), with 100% per-position argmax
agreement.

## Files
- `model.safetensors` — weights (float32, ~1.24B params; dominated by the 50257×12288 embed/unembed).
- `config.json` — `model_type: z1t` + architecture fields.
- `tokenizer.json` / `vocab.json` / `merges.txt` / `tokenizer_config.json` — standard GPT-2 BPE.
- `mlx_z1t.py`, `run.py` — self-contained MLX inference (no JAX, no mlx-vlm needed).

## Usage — standalone (no dependencies beyond MLX)
```bash
pip install mlx numpy tiktoken safetensors
python run.py "The meaning of life is" 60
```

## Usage — via mlx-vlm
Requires the `z1t` model in mlx-vlm (`mlx_vlm/models/z1t/`):
```python
from mlx_vlm import load, generate
model, processor = load("AlazarM/Z1T-0-mlx")
print(generate(model, processor, "The meaning of life is", max_tokens=60, temperature=0.0).text)
```

## Notes
- This is a **portability** port (runs on Apple Silicon), not a throughput build. Prefill is
  ~150 µs/token on an M5 Max; single-token decode is unoptimized.
- The checkpoint ships a numerically-zero positional table (AFTConv already encodes position).
- The model is a small 4-layer research artifact; generations are short and repetitive.

Original weights © Extropic, Apache-2.0. This conversion redistributes them unmodified in MLX form.