File size: 8,900 Bytes
9216df7 f916968 9216df7 1aa7540 c0bc5c8 f6358a5 c0bc5c8 e3a1ce4 54e26e6 1aa7540 cb5b591 c0bc5c8 1afb347 fd1ec67 1afb347 c0bc5c8 ad671c4 c0bc5c8 0851fe5 ac685a1 c0bc5c8 f916968 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 |
---
license: cc-by-nc-4.0
model-index:
- name: SOLAR-math-2x10.7b-v0.2
results:
- task:
type: text-generation
name: Text Generation
dataset:
name: AI2 Reasoning Challenge (25-Shot)
type: ai2_arc
config: ARC-Challenge
split: test
args:
num_few_shot: 25
metrics:
- type: acc_norm
value: 70.9
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/SOLAR-math-2x10.7b-v0.2
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: HellaSwag (10-Shot)
type: hellaswag
split: validation
args:
num_few_shot: 10
metrics:
- type: acc_norm
value: 88.29
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/SOLAR-math-2x10.7b-v0.2
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU (5-Shot)
type: cais/mmlu
config: all
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 66.25
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/SOLAR-math-2x10.7b-v0.2
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: TruthfulQA (0-shot)
type: truthful_qa
config: multiple_choice
split: validation
args:
num_few_shot: 0
metrics:
- type: mc2
value: 71.68
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/SOLAR-math-2x10.7b-v0.2
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: Winogrande (5-shot)
type: winogrande
config: winogrande_xl
split: validation
args:
num_few_shot: 5
metrics:
- type: acc
value: 83.5
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/SOLAR-math-2x10.7b-v0.2
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GSM8k (5-shot)
type: gsm8k
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 64.9
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/SOLAR-math-2x10.7b-v0.2
name: Open LLM Leaderboard
---
# 🌞🚀 SOLAR-math-10.7x2-v0.2_19B
Merge of two Solar-10.7B instruct finetunes.
This model performs in line with GPT-3.5 and Gemini Pro. Exceeding all scores of Mixtral-8x7b
Here is a brief overview of the evaluation results. These are simply for the user to have the values available for comparison. This table does not represent a complete analysis.
## 🌅 Code Example
Example also available in [colab](https://colab.research.google.com/drive/10FWCLODU_EFclVOFOlxNYMmSiLilGMBZ?usp=sharing)
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
def generate_response(prompt):
"""
Generate a response from the model based on the input prompt.
Args:
prompt (str): Prompt for the model.
Returns:
str: The generated response from the model.
"""
# Tokenize the input prompt
inputs = tokenizer(prompt, return_tensors="pt")
# Generate output tokens
outputs = model.generate(**inputs, max_new_tokens=512, eos_token_id=tokenizer.eos_token_id, pad_token_id=tokenizer.pad_token_id)
# Decode the generated tokens to a string
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
return response
# Load the model and tokenizer
model_id = "macadeliccc/SOLAR-math-2x10.7B-v0.2"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, load_in_4bit=True)
prompt = "Explain the proof of Fermat's Last Theorem and its implications in number theory."
print("Response:")
print(generate_response(prompt), "\n")
```
**Example output:**
Explain the proof of Fermat's Last Theorem and its implications in number theory.
Fermat's Last Theorem, also known as FLT, is a famous mathematical conjecture that states "no three positive integers a, b, and c can satisfy the equation a^n + b^n = c^n for any integer value of n greater than 2." This theorem was first proposed by Pierre de Fermat in the 17th century, but its proof was only discovered in the late 20th century by Andrew Wiles.
The proof of Fermat's Last Theorem, published by Andrew Wiles in 1993 and 1994, is complex and involves several advanced mathematical concepts. The main idea behind the proof is the use of modular elliptic curves, which are algebraic curves defined by polynomial equations. Wiles introduced a new concept called the Taniyama-Shimura conjecture, which states that there is a one-to-one correspondence between certain elliptic curves over the rational numbers and certain cusp forms.
Wiles' proof of FLT is based on the assumption that the Taniyama-Shimura conjecture is true. He showed that if the Taniyama-Shimura conjecture is true, then Fermat's Last Theorem must also be true. This proof strategy is known as a "proof by contradiction." Wiles demonstrated that if FLT were false, then there would exist a counterexample to the Taniyama-Shimura conjecture. However, since the Taniyama-Shimura conjecture is believed to be true, this leads to a contradiction. Therefore, by the principle of contradiction, Fermat's Last Theorem must be true.
The implications of Fermat's Last Theorem in number theory are significant. FLT is a fundamental result in the study of integers, and its proof has led to a better understanding of various mathematical concepts. The proof of FLT has also contributed to the development of other areas of mathematics, such as algebraic geometry, representation theory, and number theory itself.
Moreover, the theorem has helped to strengthen the foundations of number theory by providing a resolution to a long-standing open problem. It has also encouraged mathematicians to explore new directions in research, as the proof of FLT has opened up new avenues for investigation in related fields.
## 🏆 Evaluations
### ARC
| Task |Version| Metric | Value | |Stderr|
|-------------|------:|--------------------|-------------|---|------|
|arc_challenge| 1|acc,none | 0.68| | |
| | |acc_stderr,none | 0.01| | |
| | |acc_norm,none | 0.72| | |
| | |acc_norm_stderr,none| 0.01| | |
| | |alias |arc_challenge| | |
Average: 71.76%
### HellaSwag
| Task |Version| Metric | Value | |Stderr|
|---------|------:|--------------------|---------|---|------|
|hellaswag| 1|acc,none | 0.71| | |
| | |acc_stderr,none | 0| | |
| | |acc_norm,none | 0.88| | |
| | |acc_norm_stderr,none| 0| | |
| | |alias |hellaswag| | |
Average: 88.01%
### 📚 Citations
```bibtex
@misc{kim2023solar,
title={SOLAR 10.7B: Scaling Large Language Models with Simple yet Effective Depth Up-Scaling},
author={Dahyun Kim and Chanjun Park and Sanghoon Kim and Wonsung Lee and Wonho Song and Yunsu Kim and Hyeonwoo Kim and Yungi Kim and Hyeonju Lee and Jihoo Kim and Changbae Ahn and Seonghoon Yang and Sukyung Lee and Hyunbyung Park and Gyoungjin Gim and Mikyoung Cha and Hwalsuk Lee and Sunghun Kim},
year={2023},
eprint={2312.15166},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_macadeliccc__SOLAR-math-2x10.7b-v0.2)
| Metric |Value|
|---------------------------------|----:|
|Avg. |74.25|
|AI2 Reasoning Challenge (25-Shot)|70.90|
|HellaSwag (10-Shot) |88.29|
|MMLU (5-Shot) |66.25|
|TruthfulQA (0-shot) |71.68|
|Winogrande (5-shot) |83.50|
|GSM8k (5-shot) |64.90|
|