Skip to content
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

Jump to bottom

Add Megatron .bin file sharding and unsharding scripts #45

Merged
merged 4 commits into from
Jan 4, 2023
Merged

Add Megatron .bin file sharding and unsharding scripts #45

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
fc2fc6b
Add .bin file sharding and unsharding scripts
haileyschoelkopf Dec 27, 2022
f3eb842
recent fixes to sharding scripts
haileyschoelkopf Dec 27, 2022
47f5541
correct size of final shard
haileyschoelkopf Dec 28, 2022
011a473
cleanup unsharding script
haileyschoelkopf Jan 1, 2023
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
Prev Previous commit
cleanup unsharding script
  • Loading branch information
@haileyschoelkopf
haileyschoelkopf committed Jan 1, 2023
commit 011a473f6dae5bfef99915074d166276a0210f62
10 changes: 0 additions & 10 deletions utils/unshard_memmap.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,20 +35,10 @@ def unshard(
print(shard_filename)
shard_memmap = np.memmap(shard_filename, mode="r", order="C")

#if full_idx_map is None:
# full_idx_map = shard_memmap
#else:
# full_idx_map = np.concatenate([full_idx_map, shard_memmap])
size = SHARD_SIZE if not (i == num_shards - 1) else final_shard_size
full_idx_map[i * SHARD_SIZE: (i * SHARD_SIZE) + size] = shard_memmap

del shard_memmap
print(full_idx_map.shape)

# write full file
with open(os.path.join(output_dir, base_filename) + ".bin", "wb+") as out_full_file:
full_idx_map.tofile(out_full_file)


if __name__ == "__main__":

Expand Down