Undi95

/

MLewd-L2-Chat-13B

like 30

Hello!
There is no dataset, it's a merge of multiple LoRA/Model

• Undi95/MLewd-L2-13B-Part3 (checkpoint of MLewd without LORA)
• posicube/Llama2-chat-AYT-13B
• zattio770/120-Days-of-LORA-v2-13B
• royallab/Pygmalion-2-13b-SuperCOT
• Undi95/MMSoul-13b-lora

For a starting point you could check :

• Pygmalion datasets
• 120-Days-of-LORA-v2-13B
• SuperCOT
• All the model/dataset used in https://huggingface.co/The-Face-Of-Goonery/Huginn-13b-FP16

Sorry!

Edit: I gave you access to the private model Undi95/ReMM-S-Light and Undi95/MLewd-L2-13B-Part3, the second one was done with an old version of mergekit, all model used to do it are public (outside ReMM-S-Light), here is the recipe:

python main.py ./MLewd-L2-13B-Part1 --cuda --method slerp --base-model Undi95/ReMM-S-Light --merge Undi95/CreativityEngine --weight 0.25
=> MLewd-L2-13B-Part1 (0.75 ReMM / 0.25 CreativeEngine)

python main.py ./MLewd-L2-13B-Part2 --cuda --method slerp --base-model Undi95/MLewd-L2-13B-Part1 --merge Sao10K/Stheno-Inverted-L2-13B --weight 0.20
=> MLewd-L2-13B-Part2 (0.64 ReMM / 0.16 CreateEngine / 0.25 Stheno-Inverted)

python main.py ./MLewd-L2-13B-Part3 --cuda --method slerp --base-model Undi95/MLewd-L2-13B-Part2 --merge The-Face-Of-Goonery/Huginn-v3-13b --weight 0.15
=> MLewd-L2-13B-Part3 (0.544 ReMM / 0.136 CreateEngine / 0.2125 Stheno-Inverted / 0.15 Huginn-v3)

Thanks for your response ! Can u share access with me to this models, pls

They're now open