What are the hardware requirements for TinyExperts V0 4x1B?

TinyExperts V0 4x1B requires approximately 5.2 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed TinyExperts V0 4x1B and how large is it?

TinyExperts V0 4x1B is developed by beberik, a model with 2.6b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does TinyExperts V0 4x1B perform on standard benchmarks?

TinyExperts V0 4x1B has the following published scores: MMLU 25.87. Compare against reference models on this page or on the LLM Explorer leaderboards.

TinyExperts V0 4x1B by beberik — VRAM 5.2GB, 32K context

Name: TinyExperts V0 4x1B
Author: beberik

TinyExperts V0 4x1B is an open-source language model by beberik. Features: 2.6b LLM, VRAM: 5.2GB, Context: 32K, License: cc-by-nc-4.0, MoE, LLM Explorer Score: 0.16, Arc: 31.4, HellaSwag: 52.3, MMLU: 25.9, GSM8K: 0.5.

Autotrain compatible Endpoints compatible Llama Merge Mixtral Model-index Moe Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/beberik/TinyExperts-v0-4x1B

TinyExperts V0 4x1B Parameters and Internals

Model Type

text-generation

Additional Notes

Some experiment with 4 different tiny llama finetunes. Claimed to be smarter than GPT-6.

LLM Name	TinyExperts V0 4x1B
Repository 🤗	https://huggingface.co/beberik/TinyExperts-v0-4x1B
Model Size	2.6b
Required VRAM	5.2 GB
Updated	2024-12-02
Maintainer	beberik
Model Type	mixtral
Model Files	5.2 GB: 1-of-1
Model Architecture	MixtralForCausalLM
License	cc-by-nc-4.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.37.0.dev0
Tokenizer Class	LlamaTokenizer
Padding Token	<s>
Vocabulary Size	32003
Torch Data Type	float16

Rank the TinyExperts V0 4x1B capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

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

Best Alternatives to TinyExperts V0 4x1B

Best Alternatives	Context / RAM	Downloads	Likes
...oE 2.6B A0.6B Instruct Preview	4K / 5.2 GB	1764	11
Aquif 3.5 A0.6B Preview	4K / 5.2 GB	17	0
EuroMoE 2.6B A0.6B 2512	4K / 5.2 GB	139	11
TinyEnsemble 3x1.1B TinyMoE	4K / 5.3 GB	6	0
...x MoE Samantha Chatml Therabot	4K / 5.2 GB	11	1
...Oasst2 ChatML Cluster 3 2 1 V1	4K / 5.2 GB	6	3
LiteLlamix 8x460M 1T	1K / 5.2 GB	9	4

Note: green Score (e.g. "73.2") means that the model is better than beberik/TinyExperts-v0-4x1B.