bhavinjawade/nectororca-solar10b-jawade
bhavinjawade/nectororca-solar10b-jawade is a 10.7 billion parameter instruction-tuned causal language model, fine-tuned by bhavinjawade from the Upstage SOLAR-10.7B-Instruct-v1.0 base model. It was trained using LoRA on the Intel DPO Orca dataset, showing slight improvements on OpenLLM Leaderboard benchmarks compared to its base. This model is primarily optimized for instruction following and general conversational tasks, leveraging its DPO training for improved alignment.
Loading preview...
SOLAR-10B-OrcaDPO-Jawade Overview
This model, bhavinjawade/nectororca-solar10b-jawade, is an instruction-tuned version of the Upstage SOLAR-10.7B-Instruct-v1.0 model, featuring 10.7 billion parameters. It was developed by bhavinjawade through fine-tuning with LoRA (Low-Rank Adaptation) on the Intel DPO Orca dataset.
Key Capabilities
- Enhanced Instruction Following: Fine-tuned specifically for better adherence to instructions.
- Improved Alignment: Leverages DPO (Direct Preference Optimization) training for more aligned and helpful responses.
- General Conversational AI: Suitable for a wide range of chat-based interactions.
- Slight Benchmark Improvements: Demonstrates minor (less than 1%) improvements on OpenLLM Leaderboard benchmarks compared to the base
SOLAR 10.7B-Instructmodel, and significant improvement overSOLAR 10.7B.
Good for
- Developers seeking an instruction-tuned model for general chat applications.
- Use cases requiring a model with improved alignment through DPO training.
- Experimentation with LoRA-based fine-tuning on a robust base model.
This model is released under the MIT License, permitting broad reuse, modification, and distribution for both private and commercial purposes.