Quantization made by Richard Erkhov.

Llama3-KALE-LM-Chem-1.5-8B - GGUF

Model creator: https://huggingface.co/USTC-KnowledgeComputingLab/
Original model: https://huggingface.co/USTC-KnowledgeComputingLab/Llama3-KALE-LM-Chem-1.5-8B/

Name	Quant method	Size
Llama3-KALE-LM-Chem-1.5-8B.Q2_K.gguf	Q2_K	2.96GB
Llama3-KALE-LM-Chem-1.5-8B.IQ3_XS.gguf	IQ3_XS	3.28GB
Llama3-KALE-LM-Chem-1.5-8B.IQ3_S.gguf	IQ3_S	3.43GB
Llama3-KALE-LM-Chem-1.5-8B.Q3_K_S.gguf	Q3_K_S	3.41GB
Llama3-KALE-LM-Chem-1.5-8B.IQ3_M.gguf	IQ3_M	3.52GB
Llama3-KALE-LM-Chem-1.5-8B.Q3_K.gguf	Q3_K	3.74GB
Llama3-KALE-LM-Chem-1.5-8B.Q3_K_M.gguf	Q3_K_M	3.74GB
Llama3-KALE-LM-Chem-1.5-8B.Q3_K_L.gguf	Q3_K_L	4.03GB
Llama3-KALE-LM-Chem-1.5-8B.IQ4_XS.gguf	IQ4_XS	4.18GB
Llama3-KALE-LM-Chem-1.5-8B.Q4_0.gguf	Q4_0	4.34GB
Llama3-KALE-LM-Chem-1.5-8B.IQ4_NL.gguf	IQ4_NL	1.49GB
Llama3-KALE-LM-Chem-1.5-8B.Q4_K_S.gguf	Q4_K_S	4.37GB
Llama3-KALE-LM-Chem-1.5-8B.Q4_K.gguf	Q4_K	4.58GB
Llama3-KALE-LM-Chem-1.5-8B.Q4_K_M.gguf	Q4_K_M	4.58GB
Llama3-KALE-LM-Chem-1.5-8B.Q4_1.gguf	Q4_1	4.78GB
Llama3-KALE-LM-Chem-1.5-8B.Q5_0.gguf	Q5_0	5.21GB
Llama3-KALE-LM-Chem-1.5-8B.Q5_K_S.gguf	Q5_K_S	5.21GB
Llama3-KALE-LM-Chem-1.5-8B.Q5_K.gguf	Q5_K	5.34GB
Llama3-KALE-LM-Chem-1.5-8B.Q5_K_M.gguf	Q5_K_M	5.34GB
Llama3-KALE-LM-Chem-1.5-8B.Q5_1.gguf	Q5_1	5.65GB
Llama3-KALE-LM-Chem-1.5-8B.Q6_K.gguf	Q6_K	6.14GB
Llama3-KALE-LM-Chem-1.5-8B.Q8_0.gguf	Q8_0	7.95GB

Original model description:

license: llama3 language: - en base_model: - meta-llama/Meta-Llama-3-8B-Instruct tags: - KALE-LM - science - chemistry pipeline_tag: text-generation

Llama3-KALE-LM-Chem-1.5-8B

Introduction

We are thrilled to present Llama3-KALE-LM-Chem-1.5-8B, a new version of our open-source KALE-LM for science, which specializes in chemistry.

We have trained our model with a larger amount of data.

Benchmarks

Open Benchmarks

Models	ChemBench	MMLU	MMLU-Chem	SciQ	IE(Acc)	IE(LS)
GPT-3.5	47.15	69.75	53.32	89.6	52.98	68.28
GPT-4	53.72	78.67	63.70	94.10	54.20	69.74
Llama3-8B-Instruct	46.02	68.3	51.10	93.30	45.83	61.22
LlaSMol	28.47	54.47	33.24	72.30	2.16	3.23
ChemDFM	44.44	58.11	45.60	86.70	7.61	11.49
ChemLLM-7B-Chat	34.16	61.79	48.39	94.00	29.66	39.17
ChemLLM-7B-Chat-1.5-SFT	42.75	63.56	49.63	95.10	14.96	19.61
Llama3-KALE-LM-Chem-1.5-8B	57.01	68.06	54.83	91.60	57.53	64.16

ChemBench Details (Evaluated By OpenCompass)

Models	NC	PP	M2C	C2M	PP	RS	YP	TP	SP	Average
GPT-3.5	46.93	56.98	85.28	38.25	43.67	42.33	30.33	42.57	38	47.15
GPT-4	54.82	65.02	92.64	52.88	62.67	52.67	42.33	24.75	35.67	53.72
Llama3-8B-Instruct	51.31	27.79	90.30	40.88	34.00	30.00	45.33	60.89	33.67	46.02
LlaSMol	27.78	29.34	31.44	23.38	25.67	24.00	37.33	34.65	22.67	28.47
ChemDFM	36.92	55.57	83.95	42.00	40.00	37.33	39.00	33.17	32.00	44.44
ChemLLM-7B-Chat	41.05	29.76	85.28	26.12	26.00	24.00	20.00	24.26	31.00	34.16
ChemLLM-7B-Chat-1.5-SFT	50.06	49.51	85.28	38.75	38.00	26.67	28.33	31.68	33.67	42.44
Llama3-KALE-LM-Chem-1.5-8B	61.33	43.44	90.30	53.62	72.67	53.67	46.00	47.03	45.00	57.01

Cite This Work

@article{dai2024kale,
  title={KALE-LM: Unleash The Power Of AI For Science Via Knowledge And Logic Enhanced Large Model},
  author={Dai, Weichen and Chen, Yezeng and Dai, Zijie and Huang, Zhijie and Liu, Yubo and Pan, Yixuan and Song, Baiyang and Zhong, Chengli and Li, Xinhe and Wang, Zeyu and others},
  journal={arXiv preprint arXiv:2409.18695},
  year={2024}
}