Nemotron-4 340B — reviews, specs & pricing
Nvidia's large open model designed for synthetic data generation.
Summary
Nemotron-4 340B is a dense 340B parameter model optimized to generate high-quality synthetic training data for other LLMs. It's released under the NVIDIA Open Model License. It also functions as a capable general chat/instruct model.
Sample use case
Used for generating synthetic fine-tuning datasets and as a reward model in RLHF pipelines. Also usable as a general instruction-following model.
Specifications
- Provider: nvidia
- License: open
- Parameters: 340B
- Context: 4k tokens
- Released: 2024-06-14
Pros
- Permissive open license
- Strong for synthetic data generation
- Reward model variant available
Cons
- Short context window
- Very large compute requirement
- Less competitive on general chat vs newer models
Average rating 0.0 from 0 community reviews on Reviuws.