This outputs coherent words sometimes. It's the result of an attempt to merge Llama-2 and Yi-6B. Coherent words are a victory but it seems resistant to further fine tuning with QLORA and I'm not inclined to spend the GPU resources required for a full fine tune.
zombyi
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the passthrough merge method using as a base.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
slices:
- sources:
- model: reallad/llamayialpaca
layer_range: [0,32]
- sources:
- model: reallad/lesslobollama2
layer_range: [0,32]
merge_method: passthrough
dtype: bfloat16
base_model: reallad/llamayialpaca
tokenizer_source: reallad/llamayialpaca