Repository Issue Activity (beta feature)

#

#3795 torch.distributed.DistNetworkError

#

#

#

#3968 loading weight is so slow for fsdp2

#

#

#4010 Ram explodes when training a very simple model with accelerate tpu

#

#

#

#3979 [Bug] FSDP2 mixed-precision upcast to fp32 is a silent no-op since v1.13.0

#3960 [Bug] RNG states from multiple backends (e.g. CUDA + HPU) are saved but only one is restored on load_state

#

#

#3964 Security: Unsafe torch.load in FSDP2 scaler path (CWE-502)

#

#3967 [Bug] prepare() silently double-wraps models and double-serializes checkpoints when called twice on the same object

#

#

#

#

#3708 Detected kernel version 5.4.250, which is below the recommended minimum of 5.5.0;

#