【Bug已解决】GRPO + vLLM corrupts SmolVLM multimodal prompts from pre-expanded image tokens 解决方案
【Bug已解决】GRPO vLLM corrupts SmolVLM multimodal prompts from pre-expanded image tokens 解决方案原始报错:GRPO vLLM corrupts SmolVLM multimodal prompts from pre-expanded image tokens
场景:用 GRPO(组相对策略优化࿰…
2026/7/20 18:46:53