ConicCat
/

Qwen3.5-Antirep-27B

Model card Files Files and versions

ConicCat/Qwen3.5-Antirep-27B

A finetune of Qwen3.5-Antirep-27B to reduce infinite looping and repetition in multiturn conversations.

Trained using int8 LoRA with DPO on 1x A100.

Downloads last month: 7

Safetensors

Model size

27B params

Tensor type

BF16

·

Inference Providers NEW

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

Model tree for ConicCat/Qwen3.5-Antirep-27B

Base model

Qwen/Qwen3.5-27B

Finetuned

(300)

this model

Quantizations

Dataset used to train ConicCat/Qwen3.5-Antirep-27B

Collection including ConicCat/Qwen3.5-Antirep-27B

Enthusiast Models

Models for 16GB+ VRAM • 4 items • Updated Mar 12