RichardErkhov/GeneZC_-_MiniMA-3B-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

MiniMA-3B - GGUF

• Model creator: https://huggingface.co/GeneZC/
• Original model: https://huggingface.co/GeneZC/MiniMA-3B/

Name	Quant method	Size
MiniMA-3B.Q2_K.gguf	Q2_K	1.09GB
MiniMA-3B.IQ3_XS.gguf	IQ3_XS	1.21GB
MiniMA-3B.IQ3_S.gguf	IQ3_S	1.27GB
MiniMA-3B.Q3_K_S.gguf	Q3_K_S	1.27GB
MiniMA-3B.IQ3_M.gguf	IQ3_M	1.33GB
MiniMA-3B.Q3_K.gguf	Q3_K	1.4GB
MiniMA-3B.Q3_K_M.gguf	Q3_K_M	1.4GB
MiniMA-3B.Q3_K_L.gguf	Q3_K_L	1.52GB
MiniMA-3B.IQ4_XS.gguf	IQ4_XS	1.55GB
MiniMA-3B.Q4_0.gguf	Q4_0	1.62GB
MiniMA-3B.IQ4_NL.gguf	IQ4_NL	1.63GB
MiniMA-3B.Q4_K_S.gguf	Q4_K_S	1.63GB
MiniMA-3B.Q4_K.gguf	Q4_K	1.72GB
MiniMA-3B.Q4_K_M.gguf	Q4_K_M	1.72GB
MiniMA-3B.Q4_1.gguf	Q4_1	1.79GB
MiniMA-3B.Q5_0.gguf	Q5_0	1.95GB
MiniMA-3B.Q5_K_S.gguf	Q5_K_S	1.95GB
MiniMA-3B.Q5_K.gguf	Q5_K	2.01GB
MiniMA-3B.Q5_K_M.gguf	Q5_K_M	2.01GB
MiniMA-3B.Q5_1.gguf	Q5_1	2.12GB
MiniMA-3B.Q6_K.gguf	Q6_K	2.31GB
MiniMA-3B.Q8_0.gguf	Q8_0	2.99GB

Original model description:

license: apache-2.0 datasets: - EleutherAI/pile - togethercomputer/RedPajama-Data-1T - p208p2002/wudao language: - en - zh library_name: transformers widget: - text: " 4 + 3 ="

MiniMA-3B

📑 arXiv | 👻 GitHub | 🤗 HuggingFace-MiniMA | 🤗 HuggingFace-MiniChat | 🤗 HuggingFace-MiniChat-1.5 | 🤖 ModelScope-MiniMA | 🤖 ModelScope-MiniChat

🆕 Updates: MiniChat-1.5-3B

❗ Must comply with LICENSE of LLaMA2 since it is derived from LLaMA2.

A language model distilled from an adapted version of LLaMA2-7B following "Towards the Law of Capacity Gap in Distilling Language Models".

Establishing a new compute-performance pareto frontier.

The following is an example code snippet to use MiniMA-3B:

import torch

from transformers import AutoModelForCausalLM, AutoTokenizer

# MiniMA
tokenizer = AutoTokenizer.from_pretrained("GeneZC/MiniMA-3B", use_fast=False)
# GPU.
model = AutoModelForCausalLM.from_pretrained("GeneZC/MiniMA-3B", use_cache=True, device_map="auto", torch_dtype=torch.float16).eval()
# CPU.
# model = AutoModelForCausalLM.from_pretrained("GeneZC/MiniMA-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

@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

Detailed results can be found here

Metric	Value
Avg.	36.2
ARC (25-shot)	43.43
HellaSwag (10-shot)	68.06
MMLU (5-shot)	28.69
TruthfulQA (0-shot)	39.76
Winogrande (5-shot)	65.98
GSM8K (5-shot)	2.73
DROP (3-shot)	4.72