Open LzhinFdu opened 5 months ago
Well, After making the input sequence length divisible by world_size * block_size, it can run normally.
Well, After making the input sequence length divisible by world_size * block_size, it can run normally.
What is block_size?
Well, After making the input sequence length divisible by world_size * block_size, it can run normally.
What is block_size?
the block_size for flash-attn
Well, After making the input sequence length divisible by world_size * block_size, it can run normally.
What is block_size?
the block_size for flash-attn
I'm sorry I don't understand. I didn't find any block_size
parameter in this repo. Could you please tell me where is it?
Well, After making the input sequence length divisible by world_size * block_size, it can run normally.
What is block_size?
the block_size for flash-attn
I'm sorry I don't understand. I didn't find any
block_size
parameter in this repo. Could you please tell me where is it?
seems here.
Can you provide a script to run dist_flash_attn? I tried setting parallel_mode to dist_flash_attn but it didn't work successfully.
When trying to use 'dist_flash_attn' with 2*A100, process 0 is stuck in torch.cuda.synchronize() of _lightseq_forward of a certain decoderlayer, while process 1 runs to this step of the next decoderlayer. Strangely, the model gets stuck on the second sample. What might be causing this bug? Is there any way to solve this problem?
https://github.com/jzhang38/EasyContext/blob/41324ec4213ad1683de7d174ad502ac1b0e51a0a/easy_context/dist_flash_attn/lightseq_async_attn.py#L291
It seems that communication of process 0 in maybe_send_recv_fwd_qkvo is not completed.