MetaMath 70B
MetaMath 70B is released 2023-10-27 in the MetaMath family with open-source; evaluate it while provider pricing coverage matures.
- Family
- MetaMath
- Released
- 2023-10-27
- Parameters
- 70B
- Architecture
- Decoder Only
- Specialization
- general
- Openness
- Open source
- License
- Apache 2.0OSI-approvedCommercial use: permitted
- Weights
- Unknown
- Code
- Unknown
- Training
- Fine-tuned
No tracked provider pricing is available yet.
About
MetaMath-70B is a large language model based on the LLaMA-2 architecture, optimized for mathematical reasoning. It leverages a 4096 context length and requires 138 GB of VRAM for inference. The model's training on the MetaMathQA dataset, which bootstraps mathematical questions, significantly enhances its problem-solving capabilities, achieving an accuracy of 82.3% on benchmarks like GSM8K. Despite its strengths, MetaMath-70B's scalability is limited by computational resource constraints during fine-tuning with QLoRA, and its proficiency is primarily in English. Further research is needed to explore its potential biases and limitations.
Providers
No active provider routes are tracked for this model yet.
Capabilities
No model capability flags are currently sourced.
Benchmark peer barsfor Coding
No task-mapped benchmark peers are available for this model yet.
Migration checks
No linked migration route is available for this model yet.