Lemma
Collection
A lemma is "something assumed" — an intermediate theorem on the path to a larger proof, or a heading that signals the subject of what follows. • 18 items • Updated
Image-Text-to-Text
MLX
Safetensors
gemma4
lemma
4bit
apple-silicon
multimodal
on-device
conversational
How to use from
PiConfigure the model in Pi
# Install Pi:
npm install -g @mariozechner/pi-coding-agent# Add to ~/.pi/agent/models.json:
{
"providers": {
"mlx-lm": {
"baseUrl": "http://localhost:8080/v1",
"api": "openai-completions",
"apiKey": "none",
"models": [
{
"id": "lthn/lemrd-mlx"
}
]
}
}
}Run Pi
# Start Pi in your project directory:
piQuick Links
Lemrd — Gemma 4 31B Dense — MLX 4-bit
The largest dense member of the Lemma model family by Lethean. An EUPL-1.2 fork of Gemma 4 31B with the Lethean Ethical Kernel (LEK) merged into the weights.
This repo hosts the MLX 4-bit build for native Apple Silicon inference via mlx-lm and mlx-vlm. For the GGUF playground (Ollama, llama.cpp) see lthn/lemrd. For the unmodified Google base see LetheanNetwork/lemrd.
Family
| Repo | Format | Bits |
|---|---|---|
lthn/lemrd |
GGUF multi-quant | Q4_K_M → BF16 |
lthn/lemrd-mlx |
MLX | 4-bit |
lthn/lemrd-mlx-8bit |
MLX | 8-bit |
lthn/lemrd-mlx-bf16 |
MLX | bf16 |
License
EUPL-1.2. See Gemma Terms of Use for upstream base model terms.
- Downloads last month
- 9
Model size
8B params
Tensor type
BF16
·
U32 ·
Hardware compatibility
Log In to add your hardware
Quantized
Model tree for lthn/lemrd-mlx
Base model
google/gemma-4-31B Finetuned
google/gemma-4-31B-it Quantized
LetheanNetwork/lemrd Finetuned
lthn/lemrd
Start the MLX server
# Install MLX LM: uv tool install mlx-lm# Start a local OpenAI-compatible server: mlx_lm.server --model "lthn/lemrd-mlx"