• Baber Abbasi's avatar
    hotfix #2262 (#2264) · 928e8bb6
    Baber Abbasi authored
    * max_length - 1 (generation always >= 1)
    
    * vllm: fix rolling prefix_token
    
    * nit: add comment
    
    * fixup! max_length should be handled for logliklihoods
    
    * Revert "fixup! max_length should be handled for logliklihoods"
    
    This reverts commit 432d1a3b754c117c3a54ea2fe792ab3a1bd09ed3.
    928e8bb6
api_models.py 24.8 KB