Skip to content

Fix attention backward dropout #81

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
shubhamchandak94 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix attention backward dropout #81

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 25 additions & 7 deletions src/nki_samples/reference/attention.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -895,13 +895,16 @@ def _flash_attn_bwd_core(
# Dropout
#####################################################################
if dropout_p > 0.0:
softmax_y_dropped = nl.ndarray((par_dim(q_seq_tile_size), k_seq_tile_size), dtype=kernel_dtype, buffer=nl.sbuf)
offset = local_i_k_seq_tile + local_i_q_seq_tile * k_seq_n_tiles \
+ head_id * k_seq_n_tiles * q_seq_n_tiles \
+ batch_id * nheads * k_seq_n_tiles * q_seq_n_tiles
offset_seed = nl.add(seed_local[0, 0], offset, mask=mask)
nl.random_seed(seed=offset_seed, mask=mask)
softmax_y[:, :] = nl.dropout(softmax_y[:, :], rate=dropout_p_local[:, 0], mask=mask)
softmax_y[:, :] = nl.multiply(softmax_y[:, :], 1 / (1 - dropout_p), mask=mask)
softmax_y_dropped[:, :] = nl.dropout(softmax_y[:, :], rate=dropout_p_local[:, 0], mask=mask)
softmax_y_dropped[:, :] = nl.multiply(softmax_y_dropped[:, :], 1 / (1 - dropout_p), mask=mask)
else:
softmax_y_dropped = softmax_y

#####################################################################
# Step 3.1 Calculate the backward gradients dL/dV, where y=softmax@V
Expand All @@ -911,24 +914,39 @@ def _flash_attn_bwd_core(
trans_dy = nisa.nc_transpose(dy_local[i_d_head_tile, :, :],
mask=mask)
dv_psum[i_d_head_tile, :, :] += \
nisa.nc_matmul(trans_dy, softmax_y[:, :], mask=mask)
nisa.nc_matmul(trans_dy, softmax_y_dropped[:, :], mask=mask)

#####################################################################
# Step 3.2 Calculate the backward gradients dL/dsoftmax, where y=softmax@V
# in value projection with matmul(stationary=dy, moving=v)
#####################################################################
softmax_dy_psum = nl.zeros((par_dim(q_seq_tile_size), k_seq_tile_size),
softmax_dy_dropped_psum = nl.zeros((par_dim(q_seq_tile_size), k_seq_tile_size),
dtype=np.float32, buffer=nl.psum)
for i_d_head_tile in nl.affine_range(d_head_n_tiles):
softmax_dy_psum[:, :] += \
softmax_dy_dropped_psum[:, :] += \
nisa.nc_matmul(dy_local[i_d_head_tile, :, :],
v_local[i_d_head_tile, :, :],
mask=mask)

softmax_dy = nl.ndarray((par_dim(q_seq_tile_size), k_seq_tile_size), dtype=kernel_dtype, buffer=nl.sbuf)
softmax_dy[:, :] = nl.copy(softmax_dy_psum[:, :], dtype=kernel_dtype,
softmax_dy_dropped = nl.ndarray((par_dim(q_seq_tile_size), k_seq_tile_size), dtype=kernel_dtype, buffer=nl.sbuf)
softmax_dy_dropped[:, :] = nl.copy(softmax_dy_dropped_psum[:, :], dtype=kernel_dtype,
mask=mask)

#####################################################################
# Step 3.3 Apply Dropout to softmax_dy_dropped
#####################################################################
if dropout_p > 0.0:
softmax_dy = nl.ndarray((par_dim(q_seq_tile_size), k_seq_tile_size), dtype=kernel_dtype, buffer=nl.sbuf)
offset = local_i_k_seq_tile + local_i_q_seq_tile * k_seq_n_tiles \
+ head_id * k_seq_n_tiles * q_seq_n_tiles \
+ batch_id * nheads * k_seq_n_tiles * q_seq_n_tiles
offset_seed = nl.add(seed_local[0, 0], offset, mask=mask)
nl.random_seed(seed=offset_seed, mask=mask)
softmax_dy[:, :] = nl.dropout(softmax_dy_dropped[:, :], rate=dropout_p_local[:, 0], mask=mask)
softmax_dy[:, :] = nl.multiply(softmax_dy[:, :], 1 / (1 - dropout_p), mask=mask)
else:
softmax_dy = softmax_dy_dropped

#####################################################################
# Step 4 Calculate the softmax backward gradients dL/dx, where y=softmax(x)
# dL/dx = y * (dL/dy - rowsum(dO_O)), where y = softmax(x)
Expand Down
Loading