Credits to Gazingstars123 for the detailed documentation about the expansion.

The model is provided as it is, no further updates are planned.

If you want to make it compatible with the inference engine, copy the missing keys from the previous complete layer that is closest to it.

In its current form, the output should be exactly the same as the 2B model.

This is a dual-encoder model, the expansion layers take the output of the Qwen3.8-4B multilingual model.

Downloads last month
-
Safetensors
Model size
4B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for nightknocker/Anima-3.6B-Qwen3.8-4B-diffusers

Finetuned
Qwen/Qwen3.5-4B
Finetuned
(2)
this model