RichardErkhov/GeneZC_-_MiniChat-3B-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

MiniChat-3B - GGUF

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

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

Original model description:

license: apache-2.0 language: - en - zh library_name: transformers widget: - text: " [|User|] Hi 👋 [|Assistant|]"

MiniChat-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 and finetuned from an adapted version of LLaMA2-7B following "Towards the Law of Capacity Gap in Distilling Language Models".

Outperforming a wide range of 3B competitors in GPT4 evaluation and even competing with several 7B chat models.

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

import torch

from transformers import AutoModelForCausalLM, AutoTokenizer

from conversation import get_default_conv_template

# MiniChat
tokenizer = AutoTokenizer.from_pretrained("GeneZC/MiniChat-3B", use_fast=False)
# GPU.
model = AutoModelForCausalLM.from_pretrained("GeneZC/MiniChat-3B", use_cache=True, device_map="auto", torch_dtype=torch.float16).eval()
# CPU.
# model = AutoModelForCausalLM.from_pretrained("GeneZC/MiniChat-3B", use_cache=True, device_map="cpu", torch_dtype=torch.float32).eval()
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
conv = get_default_conv_template("minichat")

question = "Implement a program to find the common elements in two arrays without using any extra data structures."
conv.append_message(conv.roles[0], question)
conv.append_message(conv.roles[1], None)
prompt = conv.get_prompt()
input_ids = tokenizer([prompt]).input_ids
output_ids = model.generate(
    torch.as_tensor(input_ids).to(device),
    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: "def common_elements(arr1, arr2):\n    if len(arr1) == 0:\n        return []\n    if len(arr2) == 0:\n        return arr1\n\n    common_elements = []\n    for element in arr1:\n        if element in arr2:\n            common_elements.append(element)\n\n    return common_elements"
# Multiturn conversation could be realized by continuously appending questions to `conv`.

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.	42.94
ARC (25-shot)	44.03
HellaSwag (10-shot)	67.19
MMLU (5-shot)	39.17
TruthfulQA (0-shot)	45.67
Winogrande (5-shot)	65.27
GSM8K (5-shot)	10.54
DROP (3-shot)	28.73