SOLAR-tail-10.7B-Merge-v1.0
Model Details
Model Developers Kyujin Han (kyujinpy)
Method
Using Mergekit.
Merge config
slices:
- sources:
- model: upstage/SOLAR-10.7B-v1.0
layer_range: [0, 48]
- model: Yhyu13/LMCocktail-10.7B-v1
layer_range: [0, 48]
merge_method: slerp
base_model: upstage/SOLAR-10.7B-v1.0
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 # fallback for rest of tensors
tokenizer_source: union
dtype: float16
Model Benchmark
Open Ko leaderboard
- Follow up as Ko-link.
Model | Average | ARC | HellaSwag | MMLU | TruthfulQA | Ko-CommonGenV2 |
---|---|---|---|---|---|---|
PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0 | 48.32 | 45.73 | 56.97 | 38.77 | 38.75 | 61.16 |
jjourney1125/M-SOLAR-10.7B-v1.0 | 55.15 | 49.57 | 60.12 | 54.60 | 49.23 | 62.22 |
- Follow up as En-link.
Model Average ARC HellaSwag MMLU TruthfulQA Winogrande GSM8K PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0 71.68 66.13 86.54 66.52 60.57 84.77 65.58 kyujinpy/Sakura-SOLAR-Instruct 74.40 70.99 88.42 66.33 71.79 83.66 65.20
lm-evaluation-harness
gpt2 (pretrained=PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0), limit: None, provide_description: False, num_fewshot: 0, batch_size: None
| Task |Version| Metric |Value | |Stderr|
|----------------|------:|--------|-----:|---|-----:|
|kobest_boolq | 0|acc |0.5021|Β± |0.0133|
| | |macro_f1|0.3343|Β± |0.0059|
|kobest_copa | 0|acc |0.6220|Β± |0.0153|
| | |macro_f1|0.6217|Β± |0.0154|
|kobest_hellaswag| 0|acc |0.4380|Β± |0.0222|
| | |acc_norm|0.5380|Β± |0.0223|
| | |macro_f1|0.4366|Β± |0.0222|
|kobest_sentineg | 0|acc |0.4962|Β± |0.0251|
| | |macro_f1|0.3316|Β± |0.0113|
Implementation Code
### KO-Platypus
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
repo = "PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0"
OpenOrca = AutoModelForCausalLM.from_pretrained(
repo,
return_dict=True,
torch_dtype=torch.float16,
device_map='auto'
)
OpenOrca_tokenizer = AutoTokenizer.from_pretrained(repo)
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
Metric | Value |
---|---|
Avg. | 71.68 |
AI2 Reasoning Challenge (25-Shot) | 66.13 |
HellaSwag (10-Shot) | 86.54 |
MMLU (5-Shot) | 66.52 |
TruthfulQA (0-shot) | 60.57 |
Winogrande (5-shot) | 84.77 |
GSM8k (5-shot) | 65.58 |
- Downloads last month
- 3,856
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0
Spaces using PracticeLLM/SOLAR-tail-10.7B-Merge-v1.0 5
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard66.130
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard86.540
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard66.520
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard60.570
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard84.770
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard65.580