Vicuna 7B V1.5 By lmsys: Benchmarks, Features and Detailed Analysis. Insights on Vicuna 7B V1.5.

Arxiv:2306.05685 Arxiv:2307.09288 Autotrain compatible Llama Pytorch Region:us Sharded

Vicuna 7B V1.5 Benchmarks

LMSys ELO: 1005 vs 1158 (GPT-4) ^-13.2%

ARC: 74.1^-23.1%

HellaSwag: 72.3^-24.1%

MMLU: 51^-41%

TruthfulQA: 50.33^-14.7%

WinoGrande: 72.06^-17.6%

GSM8K: 23.7^-74.2%

OpenCompass Avg: 47.1 vs 73.3 (GPT-4) ^-35.7%

Exam: 53^-31.3%

Language: 35.9^-42.1%

Knowledge: 51.8^-29.5%

Understanding: 62.1^-11.3%

Reasoning: 41.9^-43.7%

^nn.n% — How the model compares to the GPT-4.

Vicuna 7B V1.5 Parameters and Internals

LLM Name	Vicuna 7B V1.5
Repository	Open on 🤗
Model Size	7b
Required VRAM	13.5 GB
Updated	2024-07-26
Maintainer	lmsys
Model Type	llama
Model Files	10.0 GB: 1-of-2 3.5 GB: 2-of-2
Model Architecture	LlamaForCausalLM
License	llama2
Context Length	4096
Model Max Length	4096
Transformers Version	4.31.0
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	float16

Quantized Models of the Vicuna 7B V1.5

Model	Likes	Downloads	VRAM
Vicuna 7B V1.5 GGUF	15	1280	2 GB
Vicuna 7B V1.5 AWQ	4	60	3 GB
Vicuna 7B V1.5 GPTQ	15	103	3 GB
Vicuna 7B V1.5 GGML	14	10	2 GB

Best Alternatives to Vicuna 7B V1.5

Best Alternatives	HF Rank	Context/RAM	Downloads	Likes
...1M 1000000ctx AEZAKMI 3 1 1702	0.2	1024K / 13.5 GB	312	1
Yarn Llama 2 7B 128K	0.2	128K / 13.5 GB	12833	38
LLaMA 7B PoSE YaRN 128K	0.2	128K / 13.5 GB	27	2
LLaMA 7B PoSE YaRN 96K	0.2	96K / 13.5 GB	10	1
LLaMA 7B PoSE Linear 96K	0.2	96K / 27 GB	7	1
Chat Llama2 7B 80K	0.2	80K / 13.8 GB	10	0
Llama2 7B 80K	0.2	80K / 13.8 GB	7	0
Lloma Step400	0.3	64K / 13.5 GB	306	0
Lloma Step200	0.3	64K / 13.5 GB	313	0
Lloma Step50	0.2	64K / 13.5 GB	303	0

Rank the Vicuna 7B V1.5 Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

What open-source LLMs or SLMs are you in search of? 34446 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Release v2024072501

Support LLM Explorer