LLM EXPLORER 65,894 MODELS INDEXED

MiniMind MoE 198M A64M Mixed SFT V1.0 by caspar

By caspar · 270 downloads

MiniMind MoE 198M A64M Mixed SFT V1.0 is an open-source language model by caspar. Features: 198.4m LLM, VRAM: 0.4GB, Context: 32K, License: apache-2.0, MoE, LLM Explorer Score: 0.31.

  Conversational   Custom code Dataset:huggingfacefw/fineweb-... Dataset:huggingfaceh4/ultracha...   Dataset:huggingfacetb/smoltalk   Dataset:openai/gsm8k   En   Experimental   Minimind   Moe   Region:us   Safetensors

MiniMind MoE 198M A64M Mixed SFT V1.0 Parameters and Internals

LLM NameMiniMind MoE 198M A64M Mixed SFT V1.0
Repository πŸ€—https://huggingface.co/caspar/MiniMind-MoE-198M-A64M-Mixed-SFT-v1.0 
Model Size198.4m
Required VRAM0.4 GB
Updated2026-10-27
Maintainercaspar
Model Typeminimind
Model Files  0.4 GB
Supported Languagesen
Model ArchitectureMiniMindForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.57.6
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<|endoftext|>
Vocabulary Size6400