- Notifications
You must be signed in to change notification settings - Fork 4.5k
Pull requests: hpcaitech/ColossalAI
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix: remove hardcoded bfloat16 dtype in OddsRatioLoss
#6406 opened Feb 11, 2026 by Mr-Neutr0n Loading…
Fix stage_indices not initialized in get_held_layers of deepseek
#6398 opened Jan 12, 2026 by honglyua-il Loading…
6 of 11 tasks
Fix FP8 linear layer dimension check to prevent runtime error
#6393 opened Nov 12, 2025 by ssam18 Loading…
fix: wrong dp-rank condition when enable pp
#6374 opened Jul 30, 2025 by liuqh16 Loading…
2 of 11 tasks
[checkpoint_io] Fix gather_state_dict_fast
#6333 opened May 28, 2025 by pbelevich Loading…
1 of 2 tasks
[DOC]: Update the documentation of ShardConfig for 1D, 2D, 2.5D, 3D tensor parallelism
#6278 opened Apr 26, 2025 by tongyu0924 Loading…
11 tasks
[applications/ColossalChat/examples/training_scripts/lora_finetune.py]: Fixed bug, added save_interval and added auto resume functions
#6223 opened Feb 26, 2025 by bbbolt Loading…
1 task
[checkpointio]support distributed checkpoint io for model saving.
#6181 opened Jan 16, 2025 by flybird11111 Loading…
11 tasks
[fix] launching error on special env variables (#6032)
#6173 opened Dec 31, 2024 by GCS-ZHN Loading…
9 of 10 tasks
[enhance] make input datatype ready for allgather
#6162 opened Dec 18, 2024 by BurkeHulk Loading…
11 tasks
[feature] support Gemma2Model for tensor parallem training
#6122 opened Nov 9, 2024 by jing-4369 Loading…
7 of 11 tasks
[Colossalai-Ascend] Support llama2-7b, chatglm2-6b finetune and inference on NPU
#6118 opened Nov 8, 2024 by duanjunwen Loading…
11 tasks
It is recommended to use np.asarray instead of np.array to avoid cessary copies of the data
#6015 opened Aug 17, 2024 by monster29000 Loading…
11 tasks
Previous Next
ProTip! What’s not been updated in a month: updated:<2026-02-25.