Didact, release candidate 3

Work in progress. Teaching the model the basics of roleplaying so I can DPO it.

Probably bad, though tbh I don't hate it so far.

There's an mxfp4. More details when I can. Note: Don't use this one, use this one instead.

Use with mlx

pip install mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("the-fall-of-man/didact-v1rc3-mxfp4")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)
Downloads last month
9
Safetensors
Model size
21B params
Tensor type
BF16
·
U32
·
U8
·
MLX
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for the-fall-of-man/didact-v1rc3-mxfp4

Quantized
(7)
this model