Configuration Parsing Warning:In config.json: "quantization_config.bits" must be an integer

llmfan46 / Qwen3.5-27B-ultra-uncensored-heretic-v2

QUANTIZED BY: UnstableLlama

Information

2.06bpw exl3 quantization of Qwen3.5-27B-ultra-uncensored-heretic-v2 via exllamav3.
repo generated automatically with ezexl3.

Repo Data

REVISION	GiB	KL DIV	PPL
2.06bpw	9.8498	0.1342	7.6050
3.00bpw	12.67	0.0353	7.2376
4.00bpw	15.50	0.0107	6.9668
6.00bpw	21.17	0.0011	6.9606
8.00bpw	27.13	0.0004	6.9698
bf16	50.96	0.0000	6.9692

CLI Download

hf download UnstableLlama/Qwen3.5-27B-ultra-uncensored-heretic-v2-exl3-2.06bpw --local-dir ./Qwen3.5-27B-ultra-uncensored-heretic-v2-exl3-2.06bpw

Safetensors

Model size

5B params

Tensor type

F16

I16

BF16

Inference Providers NEW

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

Model tree for UnstableLlama/Qwen3.5-27B-ultra-uncensored-heretic-v2-exl3-2.06bpw

Base model

Finetuned

Quantized

(8)

this model