AIS Nemotron Reasoning+Code TIES 32B

Part of AIS Gamma Alpha Project — merging 70+ models into one universal AI.

Merge

  • Method: TIES merge via mergekit
  • Base: nvidia/OpenReasoning-Nemotron-32B
  • Donor: nvidia/OpenCodeReasoning-Nemotron-32B
  • Architecture: Qwen2ForCausalLM
  • Size: 62 GB BF16 (14 safetensors)
  • Smoke load: PASS

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained(
    "i1knower/AIS-Gamma-Nemotron-Reasoning-Code-TIES-32B", trust_remote_code=True, dtype="bfloat16"
)
tokenizer = AutoTokenizer.from_pretrained("i1knower/AIS-Gamma-Nemotron-Reasoning-Code-TIES-32B")

Organization

GAMMAai | Author: i1knower


Brand / Project Identity

This artifact is part of the GAMMAai research and model-fusion program. Historical Hugging Face repository namespace references may still appear in URLs, but the active project brand is GAMMAai.

Downloads last month
59
Safetensors
Model size
33B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for GammaAGI/AIS-Gamma-Nemotron-Reasoning-Code-TIES-32B

Quantizations
1 model