LLM EXPLORER 60,277 MODELS INDEXED

Tinyllama Moe4 by aipib

By aipib · 9 downloads

Tinyllama Moe4 is an open-source language model by aipib. Features: 1.1b LLM, VRAM: 3.7GB, Context: 2K, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.12.

Base model:gardner/tinyllama-1... Base model:merge:gardner/tinyl... Base model:merge:vihangd/dopey... Base model:vihangd/dopeytinyll...   Conversational   Endpoints compatible   Frankenmoe Gardner/tinyllama-1.1b-instruc...   Instruct   Lazymergekit   Merge   Mergekit   Mixtral   Moe   Region:us   Safetensors   Sharded   Tensorflow   Vihangd/dopeytinyllama-1.1b-v1
Model Card on HF πŸ€—: https://huggingface.co/aipib/Tinyllama-moe4 

Tinyllama Moe4 Parameters and Internals

Model Type 
mixture of experts, text generation
Additional Notes 
Tinyllama-moe4 is a Mixture of Experts (MoE) model.
LLM NameTinyllama Moe4
Repository πŸ€—https://huggingface.co/aipib/Tinyllama-moe4 
Base Model(s)  vihangd/DopeyTinyLlama-1.1B-v1   gardner/TinyLlama-1.1B-Instruct-3T   vihangd/DopeyTinyLlama-1.1B-v1   gardner/TinyLlama-1.1B-Instruct-3T
Model Size1.1b
Required VRAM3.7 GB
Updated2026-08-08
Maintaineraipib
Model Typemixtral
Instruction-BasedYes
Model Files  2.0 GB: 1-of-2   1.7 GB: 2-of-2
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length2048
Model Max Length2048
Transformers Version4.40.1
Tokenizer ClassLlamaTokenizer
Padding Token<s>
Vocabulary Size32000
Torch Data Typefloat16