dingqingy
dingqingy
I am testing a hardware that has streaming IO. From vcd waveform, it looks like that "poke" always happens at negedge. Is there a way to poke at the posedge...
Hi, I am curious about the design decision of managing both token embeddings and the final output layer at the root fsdp level instead of treating them as different layers...
As the title suggests, is torchtitan CP supported on Turing GPU? I got the error `RuntimeError: No available kernel. Aborting execution.` using the default `run_train.sh` script with CP changed to...