NEXS Qwen3 32b IF Lora

NEXS Qwen3 32b IF Lora is a LoRA adapter designed for use with the Qwen3-32B base model. The adapter is extracted using mergekit from qihoo360/Light-IF-32B and is configured for compatibility with vLLM serving environments. According to the provided instructions, it can be integrated with libraries such as PEFT by loading the Qwen/Qwen3-32B base model and then applying the adapter. The model card mentions that sanitization was applied to ensure the adapter is ready for vLLM serving by removing unsupported full-rank modules_to_save tensors, specifically from embed_tokens, lm_head, and norm layers. The tool is available on the Hugging Face platform, with usage instructions provided for integration with PEFT, as well as for deployment in environments like Google Colab and Kaggle. The evidence does not specify a particular user audience or application domain, but the adapter is presented for use in machine learning workflows that involve the Qwen3-32B model and environments that support LoRA adapters, such as vLLM. There is no information in the evidence regarding pricing, licensing, or targeted use cases beyond these technical details. Overall, NEXS Qwen3 32b IF Lora serves as a technical component for extending the functionality of the Qwen3-32B model, particularly in workflows that require LoRA adapters compatible with vLLM.