TinyMoE-200m-2x16

TinyMoE-200M-2x16

TinyMoE-200M-2x16 is a compact, highly efficient Sparse Mixture of Experts (MoE) language model built upon the Mixtral/Mistral architecture. Designed for research, edge applications, and resource-constrained environments, this model leverages an expert-routing mechanism to balance a larger total parameter capacity with ultra-low computational overhead during inference.

Model Details

• Architecture: Sparse Mixture of Experts (MoE)
• Total Parameters: ~200m total parameters
• Active Parameters per Token: ~50M active parameters
• Expert Configuration: 16 total local experts, 2 active experts routed per token (num_experts_per_tok": 2)
• Context Length: 1024 tokens
• Base Architecture: Mixtral / Mistral For Causal LM
• License: MIT

Training Data

This model was trained onFineWeb-Edu-Dedup 60% and Cosmopedia-v2 40%.

Quick Start

You can load and experiment with this model using the Hugging Face transformers library:

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "FlameF0X/TinyMoE-200m-2x16"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

input_text = "Wikipedia is a free"
inputs = tokenizer(input_text, return_tensors="pt")

outputs = model.generate(**inputs, max_new_tokens=50)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))