Wizard-Vicuna-30B-Uncensored merged with bhenrym14's airoboros-33b-gpt4-1.4.1-lxctx-PI-16384-LoRA, quantized at 4 bit.
More info about the LoRA Here. This is an alternative to SuperHOT 8k LoRA trained with LoRA_rank 64 and context extended to 16K, with airoboros 1.4.1 dataset.
It was created with GPTQ-for-LLaMA with group size 32 and act order true as parameters, to get the maximum perplexity vs FP16 model.
I HIGHLY suggest to use exllama, to evade some VRAM issues.
Use compress_pos_emb = 8 for any context up to 16384 context.
If you have 2x24 GB VRAM GPUs cards, to not get Out of Memory errors at 16384 context, use:
gpu_split: 8.8,9.2
- Downloads last month
- 17
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.