LLM EXPLORER 60,702 MODELS INDEXED

MobileLLM 1B by facebook

By facebook · 4089 downloads

MobileLLM 1B is an open-source language model by facebook. Features: 1b LLM, VRAM: 2.1GB, Context: 2K, License: cc-by-nc-4.0, LLM Explorer Score: 0.18.

  Arxiv:2402.14905   Autotrain compatible   Custom code   En   Mobilellm   Pytorch   Region:us
Model Card on HF πŸ€—: https://huggingface.co/facebook/MobileLLM-1B 

MobileLLM 1B Parameters and Internals

Model Type 
auto-regressive language model
Use Cases 
Areas:
on-device applications
Primary Use Cases:
Zero-shot common sense reasoning tasks
Additional Notes 
MobileLLM integrated several key techniques such as SwiGLU activation function, deep and thin architectures, embedding sharing, and grouped-query attention.
Training Details 
Data Sources:
Publicly available online data.
Data Volume:
1T tokens
Context Length:
2000
Hardware Used:
32 NVIDIA A100 80G GPUs
Model Architecture:
MobileLLM is an auto-regressive language model leveraging an optimized transformer architecture with techniques such as SwiGLU activation function, deep and thin architectures, embedding sharing, and grouped-query attention.
Input Output 
Input Format:
Text
Accepted Modalities:
Text
Output Format:
Text
LLM NameMobileLLM 1B
Repository πŸ€—https://huggingface.co/facebook/MobileLLM-1B 
Model Size1b
Required VRAM2.1 GB
Updated2025-05-01
Maintainerfacebook
Model Typemobilellm
Model Files  2.1 GB   2.1 GB
Model ArchitectureMobileLLMForCausalLM
Licensecc-by-nc-4.0
Context Length2048
Model Max Length2048
Transformers Version4.41.2
Tokenizer ClassLlamaTokenizer
Vocabulary Size32000
Torch Data Typefloat16