-
Notifications
You must be signed in to change notification settings - Fork 552
Pull requests: PaddlePaddle/FastDeploy
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
After enabling "top_logprobs supports passing 0 and fix max_completion_tokens", an incorrect finish_reason was returned.
contributor
#2815
opened Jul 11, 2025 by
zhenwenDang
Loading…
[Bug fix]fix num_blocks_local when small size model in TP2 running mode
#2792
opened Jul 10, 2025 by
gzy19990617
Loading…
[Feature] support c16 prefix_cache in flash_attention_v3
#2766
opened Jul 9, 2025 by
lizhenyun01
Loading…
[Feature] mm and thinking model support structred output
#2749
opened Jul 8, 2025 by
kevincheng2
Loading…
Support use safetensors with paddle.MmapStorage to load model files
contributor
#2730
opened Jul 7, 2025 by
zeroRains
Loading…
2 tasks
ProTip!
Follow long discussions with comments:>50.