Qwen/Qwen1.5-72B

Qwen1.5-72B is a 72.3 billion parameter decoder-only transformer language model developed by Qwen, serving as a beta version for Qwen2. It is pretrained on extensive data, offering stable support for a 32K context length across all model sizes. This model series features an improved tokenizer for multilingual and code adaptability, making it suitable for further fine-tuning for various language generation tasks.