Improve slurm droid

This commit is contained in:
Remi Cadene 2025-03-20 14:12:46 +00:00
parent 5d184a7811
commit 65738f0a80
5 changed files with 95 additions and 60 deletions

View File

@ -39,14 +39,15 @@ python examples/port_datasets/droid_rlds/port.py \
## Port over SLURM
### 1. Port one shard per job
First, install slurm utilities from Hugging Face:
Install slurm utilities from Hugging Face:
```bash
pip install datatrove
```
Then run this script to start porting shards of the dataset:
### 1. Port one shard per job
Run this script to start porting shards of the dataset:
```bash
python examples/port_datasets/droid_rlds/slurm_port_shards.py \
--raw-dir /your/data/droid/1.0.1 \
@ -83,7 +84,7 @@ Check if your jobs are running:
squeue -u $USER`
```
You should see a list with job indices like `15125385_155` where `15125385` is the job index and `155` is the worker index. The output/print of this worker is written in real time in `/your/logs/job_name/slurm_jobs/15125385_155.out`. For instance, you can inspect the content of this file by running `less /your/logs/job_name/slurm_jobs/15125385_155.out`.
You should see a list with job indices like `15125385_155` where `15125385` is the index of the run and `155` is the worker index. The output/print of this worker is written in real time in `/your/logs/job_name/slurm_jobs/15125385_155.out`. For instance, you can inspect the content of this file by running `less /your/logs/job_name/slurm_jobs/15125385_155.out`.
Check the progression of your jobs by running:
```bash

View File

@ -307,7 +307,7 @@ def generate_lerobot_frames(tf_episode):
def port_droid(
raw_dir: Path,
repo_id: str = None,
repo_id: str,
push_to_hub: bool = False,
num_shards: int | None = None,
shard_index: int | None = None,
@ -349,11 +349,12 @@ def port_droid(
logging.info(f"Number of episodes {num_episodes}")
for episode_index, episode in enumerate(raw_dataset):
logging.info(f"{episode_index} / {num_episodes} episodes processed")
elapsed_time = time.time() - start_time
d, h, m, s = get_elapsed_time_in_days_hours_minutes_seconds(elapsed_time)
logging.info(f"It has been {d} days, {h} hours, {m} minutes, {s:.3f} seconds")
logging.info(
f"{episode_index} / {num_episodes} episodes processed (after {d} days, {h} hours, {m} minutes, {s:.3f} seconds)"
)
for frame in generate_lerobot_frames(episode):
lerobot_dataset.add_frame(frame)

View File

@ -16,6 +16,7 @@
import argparse
import logging
from pathlib import Path
import tqdm
from datatrove.executor import LocalPipelineExecutor
@ -197,7 +198,7 @@ def make_aggregate_executor(
"pipeline": [
AggregateDatasets(repo_ids, repo_id),
],
"logging_dir": str(logs_dir),
"logging_dir": str(logs_dir / job_name),
}
if slurm:
@ -235,7 +236,7 @@ def main():
)
parser.add_argument(
"--logs-dir",
type=str,
type=Path,
help="Path to logs directory for `datatrove`.",
)
parser.add_argument(

View File

@ -67,7 +67,7 @@ def make_port_executor(
"pipeline": [
PortDroidShards(raw_dir, repo_id),
],
"logging_dir": str(logs_dir),
"logging_dir": str(logs_dir / job_name),
}
if slurm:
@ -111,7 +111,7 @@ def main():
)
parser.add_argument(
"--logs-dir",
type=str,
type=Path,
help="Path to logs directory for `datatrove`.",
)
parser.add_argument(

View File

@ -12,6 +12,7 @@ from huggingface_hub.constants import REPOCARD_NAME
from examples.port_datasets.droid_rlds.port_droid import DROID_SHARDS
from lerobot.common.datasets.lerobot_dataset import CODEBASE_VERSION, LeRobotDatasetMetadata
from lerobot.common.datasets.utils import create_lerobot_dataset_card
from lerobot.common.utils.utils import init_logging
class UploadDataset(PipelineStep):
@ -23,10 +24,12 @@ class UploadDataset(PipelineStep):
tags: list | None = None,
license: str | None = "apache-2.0",
private: bool = False,
distant_repo_id: str | None = None,
**card_kwargs,
):
super().__init__()
self.repo_id = repo_id
self.distant_repo_id = self.repo_id if distant_repo_id is None else distant_repo_id
self.branch = branch
self.tags = tags
self.license = license
@ -43,96 +46,123 @@ class UploadDataset(PipelineStep):
self.create_repo()
def create_repo(self):
hub_api = HfApi()
logging.info(f"Loading meta data from {self.repo_id}...")
meta = LeRobotDatasetMetadata(self.repo_id)
logging.info(f"Creating repo {self.distant_repo_id}...")
hub_api = HfApi()
hub_api.create_repo(
repo_id=self.repo_id,
repo_id=self.distant_repo_id,
private=self.private,
repo_type="dataset",
exist_ok=True,
)
if self.branch:
hub_api.create_branch(
repo_id=self.repo_id,
repo_id=self.distant_repo_id,
branch=self.branch,
revision=self.revision,
repo_type="dataset",
exist_ok=True,
)
if not hub_api.file_exists(self.repo_id, REPOCARD_NAME, repo_type="dataset", revision=self.branch):
if not hub_api.file_exists(
self.distant_repo_id, REPOCARD_NAME, repo_type="dataset", revision=self.branch
):
card = create_lerobot_dataset_card(
tags=self.tags, dataset_info=meta.info, license=self.license, **self.card_kwargs
)
card.push_to_hub(repo_id=self.repo_id, repo_type="dataset", revision=self.branch)
card.push_to_hub(repo_id=self.distant_repo_id, repo_type="dataset", revision=self.branch)
def list_files_recursively(directory):
base_path = Path(directory)
return [str(file.relative_to(base_path)) for file in base_path.rglob("*") if file.is_file()]
meta = LeRobotDatasetMetadata(self.repo_id)
logging.info(f"Listing all local files from {self.repo_id}...")
self.file_paths = list_files_recursively(meta.root)
self.file_paths = sorted(self.file_paths)
def run(self, data=None, rank: int = 0, world_size: int = 1):
import logging
import random
import time
def create_chunks(self, lst, n):
from itertools import islice
from huggingface_hub import CommitOperationAdd, create_commit, preupload_lfs_files
it = iter(lst)
return [list(islice(it, size)) for size in [len(lst) // n + (i < len(lst) % n) for i in range(n)]]
def create_commits(self, additions):
import logging
import math
import random
import time
from huggingface_hub import create_commit
from huggingface_hub.utils import HfHubHTTPError
FILES_BETWEEN_COMMITS = 10 # noqa: N806
BASE_DELAY = 0.1 # noqa: N806
MAX_RETRIES = 12 # noqa: N806
# Split the files into smaller chunks for faster commit
# and avoiding "A commit has happened since" error
num_chunks = math.ceil(len(additions) / FILES_BETWEEN_COMMITS)
chunks = self.create_chunks(additions, num_chunks)
for chunk in chunks:
retries = 0
while True:
try:
create_commit(
self.distant_repo_id,
repo_type="dataset",
operations=chunk,
commit_message=f"DataTrove upload ({len(chunk)} files)",
revision=self.branch,
)
logging.info("create_commit completed!")
break
except HfHubHTTPError as e:
if "A commit has happened since" in e.server_message:
if retries >= MAX_RETRIES:
logging.error(f"Failed to create commit after {MAX_RETRIES=}. Giving up.")
raise e
logging.info("Commit creation race condition issue. Waiting...")
time.sleep(BASE_DELAY * 2**retries + random.uniform(0, 2))
retries += 1
else:
raise e
def run(self, data=None, rank: int = 0, world_size: int = 1):
import logging
from datasets.utils.tqdm import disable_progress_bars
from huggingface_hub import CommitOperationAdd, preupload_lfs_files
from lerobot.common.datasets.lerobot_dataset import LeRobotDatasetMetadata
from lerobot.common.utils.utils import init_logging
BASE_DELAY = 1.0 # noqa: N806
MAX_RETRIES = 24 # noqa: N806
init_logging()
disable_progress_bars()
def chunked(lst, n):
it = iter(lst)
return [list(islice(it, size)) for size in [len(lst) // n + (i < len(lst) % n) for i in range(n)]]
chunks = chunked(self.file_paths, world_size)
chunks = self.create_chunks(self.file_paths, world_size)
file_paths = chunks[rank]
if len(file_paths) == 0:
raise ValueError(file_paths)
logging.info("Pre-uploading LFS files...")
for i, path in enumerate(file_paths):
logging.info(f"{i}: {path}")
meta = LeRobotDatasetMetadata(self.repo_id)
additions = [
CommitOperationAdd(path_in_repo=path, path_or_fileobj=meta.root / path) for path in file_paths
]
logging.info(f"Uploading {','.join(file_paths)} to the hub...")
preupload_lfs_files(
repo_id=self.repo_id, repo_type="dataset", additions=additions, revision=self.branch
repo_id=self.distant_repo_id, repo_type="dataset", additions=additions, revision=self.branch
)
logging.info(f"Upload of {','.join(file_paths)} to the hub complete!")
retries = 0
while True:
try:
create_commit(
self.repo_id,
repo_type="dataset",
operations=additions,
commit_message=f"DataTrove upload ({len(additions)} files)",
revision=self.branch,
)
break
except HfHubHTTPError as e:
if "A commit has happened since" in e.server_message:
if retries >= MAX_RETRIES:
logging.error(f"Failed to create commit after {MAX_RETRIES=}. Giving up.")
raise e
logging.info("Commit creation race condition issue. Waiting...")
time.sleep(BASE_DELAY * 2**retries + random.uniform(0, 2))
retries += 1
else:
raise e
logging.info("Creating commits...")
self.create_commits(additions)
logging.info("Done!")
def make_upload_executor(
@ -142,7 +172,7 @@ def make_upload_executor(
"pipeline": [
UploadDataset(repo_id),
],
"logging_dir": str(logs_dir),
"logging_dir": str(logs_dir / job_name),
}
if slurm:
@ -180,7 +210,7 @@ def main():
)
parser.add_argument(
"--logs-dir",
type=str,
type=Path,
help="Path to logs directory for `datatrove`.",
)
parser.add_argument(
@ -209,7 +239,7 @@ def main():
parser.add_argument(
"--cpus-per-task",
type=int,
default=4,
default=8,
help="Number of cpus that each slurm worker will use.",
)
parser.add_argument(
@ -219,6 +249,8 @@ def main():
help="Memory per cpu that each worker will use.",
)
init_logging()
args = parser.parse_args()
kwargs = vars(args)
kwargs["slurm"] = kwargs.pop("slurm") == 1