β Catalog
TinyLlama v1.1 Math & Code
AvailableTinyLlamaOpen source
TinyLlama_v1.1_math_code is the math-and-code continual-pretrain variant of the TinyLlama v1.1 1.1B family, released as open weights on Hugging Face under Apache 2.0. It uses the same Llama 2 architecture and 2048-token context as TinyLlama_v1.1, with additional StarCoder and Proof-Pile data in later pretraining stages. Hugging Face created the model repo on 2024-03-09; the card itself does not print a release date.
Specifications
- License
- Open source Β· Apache-2.0
- Weights
- Downloadable
- Architecture
- Dense
- Parameters
- 1.1B
- Context window
- 2K tokens
- Max output
- β
- Knowledge cutoff
- β
- Price (in / out, $/M)
- β
- Modalities
- Text
Benchmarks
No benchmark scores recorded yet. Spotted some? Submit a correction.
Vendor-reported figures are claims until independently verified. See methodology.