SOLAR-13B-Instruct-v1.0

This is SOLAR-10.7B, but upscaled to 13B, to optimize VRAM usage of typical GPU cards (a 4bit quant fits in 12GB).

Evaluations coming soon!

This is a frankenmerge model created using mergekit.

Merge Details

Merge Method

This model was merged using the passthrough merge method.

Models Merged

The following models were included in the merge:

• upstage/SOLAR-10.7B-Instruct-v1.0

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
    - model: upstage/SOLAR-10.7B-Instruct-v1.0
      layer_range: [0, 28]
  - sources:
    - model: upstage/SOLAR-10.7B-Instruct-v1.0
      layer_range: [20, 48]
merge_method: passthrough
dtype: float16

Prompt template

The same as in SOLAR-10.7B:

<s> ### User:
{prompt}

### Assistant:
{response}</s>

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	56.65
AI2 Reasoning Challenge (25-Shot)	57.25
HellaSwag (10-Shot)	78.03
MMLU (5-Shot)	55.75
TruthfulQA (0-shot)	61.99
Winogrande (5-shot)	70.24
GSM8k (5-shot)	16.60