File size: 7,111 Bytes
4a54cb4 f3b6e28 4a54cb4 f3b6e28 a743770 f3b6e28 4a54cb4 a743770 5fcb087 a743770 fdad192 a743770 f3b6e28 |
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 |
---
language:
- en
- zh
license: apache-2.0
library_name: transformers
datasets:
- EleutherAI/pile
- togethercomputer/RedPajama-Data-1T
- p208p2002/wudao
widget:
- text: <s> 4 + 3 =
model-index:
- name: MiniMA-2-3B
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: 44.71
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=GeneZC/MiniMA-2-3B
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: 69.33
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=GeneZC/MiniMA-2-3B
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: 41.22
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=GeneZC/MiniMA-2-3B
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: 38.44
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=GeneZC/MiniMA-2-3B
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: 66.69
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=GeneZC/MiniMA-2-3B
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: 8.11
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=GeneZC/MiniMA-2-3B
name: Open LLM Leaderboard
---
## MiniMA-2-3B
📑 [arXiv](https://arxiv.org/abs/2311.07052) | 👻 [GitHub](https://github.com/GeneZC/MiniMA) | 🤗 [HuggingFace-MiniMA](https://huggingface.co/GeneZC/MiniMA-3B) | 🤗 [HuggingFace-MiniChat](https://huggingface.co/GeneZC/MiniChat-3B) | 🤖 [ModelScope-MiniMA](https://modelscope.cn/models/GeneZC/MiniMA-3B) | 🤖 [ModelScope-MiniChat](https://modelscope.cn/models/GeneZC/MiniChat-3B) | 🤗 [HuggingFace-MiniChat-1.5](https://huggingface.co/GeneZC/MiniChat-1.5-3B) | 🤗 [HuggingFace-MiniMA-2](https://huggingface.co/GeneZC/MiniMA-2-3B) | 🤗 [HuggingFace-MiniChat-2](https://huggingface.co/GeneZC/MiniChat-2-3B)
🆕 **Updates from MiniMA-3B**:
- continued from MiniMA-3B without distillation;
- better data mixture;
- more trained tokens.
❗ Must comply with LICENSE of LLaMA-2 since it is derived from LLaMA-2.
A language model continued from MiniMA-3B.
Completing the compute-performance pareto frontier together with MiniMA-3B and other arts.
<img src="./teaser_a.jpg" alt="teaser_a" width="700" />
**Standard Benchmarks**
|Method|TFLOPs|MMLU (5-shot)|CEval (5-shot)|DROP (3-shot)|HumanEval (0-shot)|BBH (3-shot)|GSM8K (8-shot)|
|--|--|--|--|--|--|--|--|
|Mamba-2.8B|4.6E9|25.58|24.74|15.72|7.32|29.37|3.49|
|ShearedLLaMA-2.7B|0.8E9|26.97|22.88|19.98|4.88|30.48|3.56|
|BTLM-3B|11.3E9|27.20|26.00|17.84|10.98|30.87|4.55|
|StableLM-3B|72.0E9|44.75|31.05|22.35|15.85|32.59|10.99|
|Qwen-1.8B|23.8E9|44.05|54.75|12.97|14.02|30.80|22.97|
|Phi-2-2.8B|159.9E9|56.74|34.03|30.74|46.95|44.13|55.42|
|LLaMA-2-7B|84.0E9|46.00|34.40|31.57|12.80|32.02|14.10|
||
|MiniMA-3B|4.0E9|28.51|28.23|22.50|10.98|31.61|8.11|
|MiniChat-3B|4.0E9|38.40|36.48|22.58|18.29|31.36|29.72|
|MiniMA-2-3B|13.4E9|40.14|44.65|23.10|14.63|31.43|8.87|
|MiniChat-2-3B|13.4E9|46.17|43.91|30.26|22.56|34.95|38.13|
The following is an example code snippet to use MiniMA-2-3B:
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
# MiniMA
tokenizer = AutoTokenizer.from_pretrained("GeneZC/MiniMA-2-3B", use_fast=False)
# GPU.
model = AutoModelForCausalLM.from_pretrained("GeneZC/MiniMA-2-3B", use_cache=True, device_map="auto", torch_dtype=torch.float16).eval()
# CPU.
# model = AutoModelForCausalLM.from_pretrained("GeneZC/MiniMA-2-3B", use_cache=True, device_map="cpu", torch_dtype=torch.float16).eval()
prompt = "Question: Sherrie tells the truth. Vernell says Sherrie tells the truth. Alexis says Vernell lies. Michaela says Alexis tells the truth. Elanor says Michaela tells the truth. Does Elanor tell the truth?\nAnswer: No\n\nQuestion: Kristian lies. Sherrie says Kristian lies. Delbert says Sherrie lies. Jerry says Delbert tells the truth. Shalonda says Jerry tells the truth. Does Shalonda tell the truth?\nAnswer: No\n\nQuestion: Vina tells the truth. Helene says Vina lies. Kandi says Helene tells the truth. Jamey says Kandi lies. Ka says Jamey lies. Does Ka tell the truth?\nAnswer: No\n\nQuestion: Christie tells the truth. Ka says Christie tells the truth. Delbert says Ka lies. Leda says Delbert tells the truth. Lorine says Leda tells the truth. Does Lorine tell the truth?\nAnswer:"
input_ids = tokenizer([prompt]).input_ids
output_ids = model.generate(
torch.as_tensor(input_ids).cuda(),
do_sample=True,
temperature=0.7,
max_new_tokens=1024,
)
output_ids = output_ids[0][len(input_ids[0]):]
output = tokenizer.decode(output_ids, skip_special_tokens=True).strip()
# output: "No"
```
## Bibtex
```bibtex
@article{zhang2023law,
title={Towards the Law of Capacity Gap in Distilling Language Models},
author={Zhang, Chen and Song, Dawei and Ye, Zheyu and Gao, Yan},
year={2023},
url={https://arxiv.org/abs/2311.07052}
}
```
# [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_GeneZC__MiniMA-2-3B)
| Metric |Value|
|---------------------------------|----:|
|Avg. |44.75|
|AI2 Reasoning Challenge (25-Shot)|44.71|
|HellaSwag (10-Shot) |69.33|
|MMLU (5-Shot) |41.22|
|TruthfulQA (0-shot) |38.44|
|Winogrande (5-shot) |66.69|
|GSM8k (5-shot) | 8.11|
|