-
Notifications
You must be signed in to change notification settings - Fork 88
Pull requests: Lightning-AI/lightning-thunder
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Update cuDNN SDPA execution functions to use the strides of attn_mask
cudnn
executors
strides
#1798
opened Feb 25, 2025 by
IvanYashchuk
Loading…
allow embedding without renorm to be lowered as nvfuser index_select
#1796
opened Feb 24, 2025 by
jjsjann123
•
Draft
Update
test_vjp_correctness_sdpa_manual
with fp16/bf16 friendly scale value
#1791
opened Feb 24, 2025 by
riccardofelluga
Loading…
add torch.optim._functional.rmsprop benchmark
#1788
opened Feb 22, 2025 by
khushi-411
Loading…
2 of 4 tasks
benchmark_litgpt - update
low_precision_mode
accepted values
#1779
opened Feb 18, 2025 by
kshitij12345
•
Draft
Remove requires_grad specification from meta functions
autograd
#1769
opened Feb 17, 2025 by
IvanYashchuk
Loading…
add torch.optim._functional.sgd benchmark
#1766
opened Feb 15, 2025 by
khushi-411
Loading…
2 of 4 tasks
Revert updates to saved_tensors info on fw_extrace, no need to remove duplicate variables
autograd
#1756
opened Feb 7, 2025 by
IvanYashchuk
Loading…
Update vjp implementation to attach residuals to all outputs
autograd
#1755
opened Feb 7, 2025 by
IvanYashchuk
•
Draft
Simplify and fix match_fw_and_bw_saved_for_bw_proxies implementation
tracing architecture
#1754
opened Feb 7, 2025 by
IvanYashchuk
•
Draft
Relax test_te_with_retain_graph xfail strictness in test_transformer_engine_executor.py
#1718
opened Jan 30, 2025 by
nWEIdia
Loading…
3 tasks
[DO NOT REVIEW][SMOKE TEST] Skip prims shape in grad transform
#1704
opened Jan 27, 2025 by
jjsjann123
•
Draft
Previous Next
ProTip!
no:milestone will show everything without a milestone.