MemTensor/MemOperator-0.6B

Hugging Face
TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.8BQuant:BF16Context Size:32kTool Calling:SupportedLicense:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

MemOperator-0.6B is a 0.6 billion parameter causal language model from the MemTensor series, fine-tuned from the Qwen3 architecture. Developed for MemOS, it specializes in memory extraction and clustering-based memory reorganization from conversations and documents. This model supports local-only deployment for memory operations at lower cost and higher speed, offering multilingual support for English and Chinese.

Loading preview...

MemOperator-0.6B: Specialized Memory Operations for MemOS

MemOperator-0.6B is a 0.6 billion parameter causal language model, part of the MemOperator series developed by MemTensor. Fine-tuned from the Qwen3 architecture using supervised fine-tuning on human-annotated and model-generated data, it is specifically designed for memory-related operations within the MemOS system. The model's primary goal is to enable local-only deployment of MemOS, ensuring efficient and high-speed memory handling with reduced resource consumption.

Key Capabilities

  • Memory Extraction: Accurately extracts high-quality memories from both conversations and documents, including summarization of document snippets.
  • Memory Reorganization: Implements clustering-based reorganization to group and integrate related memories, enhancing long-term memory coherence.
  • Multilingual Support: Supports memory extraction and instruction following in both English and Chinese.
  • Resource Efficiency: Optimized for low-resource usage, making it suitable for edge devices and low-latency applications, while maintaining strong performance.
  • Context Length: Features a context length of 32,768 tokens.

Good For

  • Local-only AI applications: Ideal for environments requiring offline memory processing.
  • Real-time memory management: Optimized for fast and accurate memory handling.
  • Integrating with MemOS: Seamlessly integrates for memory extraction and reorganization tasks.
  • Cost-effective deployment: Offers comparable memory processing performance to larger models like Qwen3-32B with significantly reduced resource requirements.

Finetunes
1 models