## Source Model Ingredients This merge combines three Llama 3.2-based components: - [NeuraLakeAi/iSA-02-Nano-Llama-3.2-1B](https://huggingface.co/NeuraLakeAi/iSA-02-Nano-Llama-3.2-1B) A Llama-3.2-1B-based base model optimized for reasoning, with an advertised 1,048,576-token context window. - [OctoThinker/OctoThinker-1B-Hybrid-Base](https://huggingface.co/OctoThinker/OctoThinker-1B-Hybrid-Base) A 1B Llama-family hybrid base model designed with mid-training insights to support reinforcement-learning-friendly behavior. - [meta-llama/Llama-3.2-1B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct) Meta’s instruction-tuned 1B multilingual Llama 3.2 model, optimized for dialogue, summarization, and agentic retrieval tasks.

Update README.md

Update README.md

Update README.md

Upload model.safetensors

Update README.md

Upload 2 files

Upload 4 files

initial commit