Downloads · 30 days
114
6% of all-time downloads
RichardErkhov/SanjiWatsuki_-_openchat-3.5-1210-starling-slerp-gguf
SanjiWatsuki_-_openchat-3.5-1210-starling-slerp-gguf is a machine learning model from RichardErkhov. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
114
6% of all-time downloads
All-time downloads
1.8K
Public
Repo size
82.2 GB
Likes
0
Public
Click a slice to open those files.
.gguf95.2 GB · 100%
From the Hugging Face model README
Quantization made by Richard Erkhov.
openchat-3.5-1210-starling-slerp - GGUF
license: cc-by-4.0 language:
This model uses the Slerp merge method from 2 models:
I SLERPed these two together because they're both OpenChat-ish models. Fundamentally, OpenChat-3.5-1210 appears to be trained similarly to OpenChat-3.5 but now with Feedback-Collection and a de-contaminated Capybara. Starling is OpenChat-3.5 but trained with a novel training method on the Nectar set.
My hope is that a SLERP between the two retains the benefits of both.
The yaml config file for this model is here:
slices:
- sources:
- model: openchat/openchat-3.5-1210
layer_range: [0, 32]
- model: berkeley-nest/Starling-LM-7B-alpha
layer_range: [0, 32]
merge_method: slerp
base_model: openchat/openchat-3.5-1210
parameters:
t:
- filter: self_attn
value: [0, 0.5, 0.3, 0.7, 1]
- filter: mlp
value: [1, 0.5, 0.7, 0.3, 0]
- value: 0.5
dtype: bfloat16