Configuration Parsing
Warning:
In UNKNOWN_FILENAME: "auto_map.AutoTokenizer" must be a string
This model is not affiliated with AI2.
AI2 did not release an OLMo 1B Instruct model, so this is a version of OLMo 1B fine-tuned only on LIMA, for use in compute-limited experiments. We used the Tulu codebase and fine-tuned for 2 epochs with lr=2e-5.
For more information and setup instructions, see the base model allenai/OLMo-1B and the instruction fine-tuned 7B variant (trained on the full Tulu data mixture) allenai/OLMo-7B-Instruct.
Benchmark scores coming soon!
- Downloads last month
- 13