LetheanNetwork/lemer-mlx-bf16

Gemma 4 E2B in MLX format, full bf16 precision, converted from LetheanNetwork/lemer's bf16 safetensors via mlx_lm.convert --dtype bfloat16. No quantization โ€” this is the full-precision reference for the MLX family. For smaller / faster variants see LetheanNetwork/lemer-mlx (4-bit) or LetheanNetwork/lemer-mlx-8bit. For the LEK-merged variant see lthn/lemer.

Variants in this family

Repo Format Bits Use case
LetheanNetwork/lemer safetensors + gguf Q4_K_M bf16 / 4 Source weights + llama.cpp/Ollama
LetheanNetwork/lemer-mlx mlx 4 Apple Silicon default
LetheanNetwork/lemer-mlx-8bit mlx 8 Higher precision
LetheanNetwork/lemer-mlx-bf16 mlx bf16 This repo โ€” full-precision reference

Usage

from mlx_lm import load, generate

model, tokenizer = load("LetheanNetwork/lemer-mlx-bf16")
response = generate(
    model, tokenizer,
    prompt=tokenizer.apply_chat_template(
        [{"role": "user", "content": "Hello"}],
        add_generation_prompt=True,
        enable_thinking=True,
    ),
    max_tokens=512,
)

Provenance

  • Source: LetheanNetwork/lemer bf16 safetensors (= google/gemma-4-E2B-it)
  • Converter: mlx_lm.convert --dtype bfloat16 (no quantization)
  • License: Apache 2.0 (Gemma Terms of Use)

License

Apache 2.0, subject to the Gemma Terms of Use.

Downloads last month
13
Safetensors
Model size
5B params
Tensor type
BF16
ยท
MLX
Hardware compatibility
Log In to add your hardware

Quantized

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for LetheanNetwork/lemer-mlx-bf16

Finetuned
(479)
this model