NicomEdu-Alpha
NicomEdu-Alpha is an experimental public language model specialized for SS2 Mathematics classroom-style responses.
Release status
Version v0.1.0-alpha is for experimentation, research, and feedback collection.
Model details
- Base model: IFM/K2-Horizon-3.7B
- Adaptation: parameter-efficient fine-tuning merged into the base-model weights
- Primary evaluated scope: SS2 Mathematics classroom-style responses
- Release format: merged
safetensorsmodel weights
Intended use
Use NicomEdu-Alpha to experiment with mathematics explanations and classroom-style working. Performance outside SS2 Mathematics was not evaluated for this release.
Limitations
- The model can make incorrect answers.
- The model can repeat text or continue after an answer is complete.
- The model can use inconsistent formatting.
- Verify important answers independently.
Do not use this Alpha release as the sole basis for high-stakes educational, legal, medical, financial, or safety decisions.
Quick start
python import torch from transformers import AutoModelForCausalLM, AutoTokenizer model_id = "nicomedu-labs/NicomEdu-Alpha" tokenizer = AutoTokenizer.from_pretrained( model_id, trust_remote_code=True, ) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True, )
Release resources
License and attribution
NicomEdu-Alpha is distributed under Apache-2.0. It is derived from IFM/K2-Horizon-3.7B. See LICENSE and NOTICE.