Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
Paper • 2203.05482 • Published • 8
This model was converted to GGUF format from v000000/Psyonic-Rose-20B-Higher-Quality using llama.cpp
Refer to the original model card for more details on the model
Psyonic-Rose 20B Q4_K_M GGUF
Speculative recreation of jebcarter Psyonic-Rose-20B (Llama2)
merge
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the linear merge method.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
models:
- model: DavidAU/Psyonic-Cetacean-V1-20B-Ultra-Quality-Float32
parameters:
weight: 1.0
- model: tavtav/Rose-20B(fp16)
parameters:
weight: 0.05
merge_method: linear
dtype: float32
- credits jebcarter
- credits DavidAU
- credtis tavtav
- credits NeverSleep
- credits CalderaAI
{{{Alpaca instruct format}}}
- Downloads last month
- 3
Hardware compatibility
Log In to add your hardware
4-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for v000000/Psyonic-Rose-20B-Higher-Quality-Q4_K_M-GGUF
Base model
tavtav/Rose-20B Finetuned
v000000/Psyonic-Rose-20B-Higher-Quality