ExllamaV2 version of the model created by Undi!
Original Model https://huggingface.co/Undi95/Lewd-Sydney-20B
Requires ExllamaV2, which is being developed by turboderp https://github.com/turboderp/exllamav2 under an MIT license.
I could load 6bpw 24gb card without cfg cash with 4096 context. 8bpw required about 30gb to load at 4096 context.
<div style="width: 100%;"> <img src="https://cdn-uploads.huggingface.co/production/uploads/63ab1241ad514ca8d1430003/ppZDyjjZJPGihhckQb5zQ.png" style="width: 40%; min-width: 200px; display: block; margin: auto;"> </div>
This model is based on Free Sydney V2, trying to get a... lewder assistant, you get it now.
<!-- description start -->
Description
This repo contain fp16 files of Lewd-Sydney-20B, an attempt to get our beloved Sydney open to R-18 content.
<!-- description end --> <!-- description start -->
Models and loras used
- Free_Sydney_V2_13b_HF
- Undi95/Xwin-MLewd-13B-V0.2
- lemonilia/LimaRP-Llama2-13B-v3-EXPERIMENT
- Synthia v1.2 private LoRA
<!-- description end --> <!-- prompt-template start -->
Prompt template: Alpaca
Below is an instruction that describes a task. Write a response that appropriately completes the request.
### Instruction:
{prompt}
### Response:
If you want to support me, you can here.