TransferQueue/tutorial/04_custom_sampler.py at dev · TransferQueue/TransferQueue · GitHub

This repository was archived by the owner on Jan 21, 2026. It is now read-only.

History

414 lines (328 loc) · 13.7 KB

Raw

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

# Licensed under the Apache License, Version 2.0 (the "License");

# you may not use this file except in compliance with the License.

# You may obtain a copy of the License at

# http://www.apache.org/licenses/LICENSE-2.0

# Unless required by applicable law or agreed to in writing, software

# distributed under the License is distributed on an "AS IS" BASIS,

# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

# See the License for the specific language governing permissions and

# limitations under the License.

import logging

import sys

import textwrap

import time

import warnings

from pathlib import Path

from typing import Any

import numpy as np

import ray

import torch

from omegaconf import OmegaConf

from tensordict import TensorDict

logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")

logger = logging.getLogger(__name__)

warnings.filterwarnings(

action="ignore",

message=r"The PyTorch API of nested tensors is in prototype stage*",

category=UserWarning,

module=r"torch\.nested",

)

warnings.filterwarnings(

action="ignore",

message=r"Tip: In future versions of Ray, Ray will no longer override accelerator visible "

r"devices env var if num_gpus=0 or num_gpus=None.*",

category=FutureWarning,

module=r"ray\._private\.worker",

)

# Add the parent directory to the path

parent_dir = Path(__file__).resolve().parent.parent

sys.path.append(str(parent_dir))

from transfer_queue import ( # noqa: E402

SimpleStorageUnit,

TransferQueueClient,

TransferQueueController,

process_zmq_server_info,

)

from transfer_queue.sampler import BaseSampler # noqa: E402

class RandomSamplerWithReplacement(BaseSampler):

"""

Sampler 1: Random Sampling with Replacement

Samples data randomly with replacement.

Useful when you want to revisit samples multiple times.

"""

def __init__(self, seed: int = None):

super().__init__()

self.seed = seed

self._states["rng"] = np.random.RandomState(seed)

def sample(

self,

ready_indexes: list[int],

batch_size: int,

*args: Any,

**kwargs: Any,

) -> tuple[list[int], list[int]]:

rng = self._states["rng"]

if len(ready_indexes) < batch_size:

raise ValueError("Not enough ready indexes to sample from.")

# Do sample

sampled_indexes = rng.choice(ready_indexes, size=batch_size, replace=False).tolist()

# Label no consumption since samples can be reused

consumed_indexes = []

return sampled_indexes, consumed_indexes

class RandomSamplerWithoutReplacement(BaseSampler):

"""

Sampler 2: Random Sampling without Replacement

Samples data randomly without replacement.

Useful for training without data ordering bias.

"""

def __init__(self, seed: int = None):

super().__init__()

self.seed = seed

self._states["rng"] = np.random.RandomState(seed)

def sample(

self,

ready_indexes: list[int],

batch_size: int,

*args: Any,

**kwargs: Any,

) -> tuple[list[int], list[int]]:

rng = self._states["rng"]

if len(ready_indexes) < batch_size:

raise ValueError("Not enough ready indexes to sample from.")

# Do sample

sampled_indexes = rng.choice(ready_indexes, size=batch_size, replace=False).tolist()

# Consumed indexes are same as sampled

consumed_indexes = sampled_indexes.copy()

return sampled_indexes, consumed_indexes

class PrioritySampler(BaseSampler):

"""

Sampler 3: Priority Sampling

Samples based on priority scores (e.g., loss, uncertainty, etc.).

Priority can be longer than ready_indexes - use partial sampling.

"""

def __init__(

self,

super().__init__()

def sample(

self,

ready_indexes: list[int],

batch_size: int,

priority_scores: np.ndarray = None,

*args: Any,

**kwargs: Any,

) -> tuple[list[int], list[int]]:

if len(ready_indexes) < batch_size:

raise ValueError("Not enough ready indexes to sample from.")

if priority_scores is None:

priority_scores = np.ones(len(ready_indexes), dtype=float)

elif len(priority_scores) > len(ready_indexes):

# Priority longer than ready_indexes - use partial

priority_scores = priority_scores[ready_indexes]

# Convert scores to probabilities

priority_scores = priority_scores / priority_scores.sum()

# Sample without replacement

sampled_indexes = np.random.choice(

ready_indexes, size=min(batch_size, len(ready_indexes)), replace=False, p=priority_scores

).tolist()

consumed_indexes = sampled_indexes.copy()

return sampled_indexes, consumed_indexes

def setup_transfer_queue_with_sampler(sampler):

"""Setup TransferQueue with custom sampler."""

if not ray.is_initialized():

ray.init()

config = OmegaConf.create(

{

"global_batch_size": 8,

"num_data_storage_units": 2,

}

)

storage_units = {}

for i in range(2):

storage_units[i] = SimpleStorageUnit.remote(storage_unit_size=100)

controller = TransferQueueController.remote(sampler=sampler)

controller_info = process_zmq_server_info(controller)

storage_unit_infos = process_zmq_server_info(storage_units)

client = TransferQueueClient(

client_id="TutorialClient",

controller_info=controller_info,

)

tq_config = OmegaConf.create({}, flags={"allow_objects": True})

tq_config.controller_info = controller_info

tq_config.storage_unit_infos = storage_unit_infos

config = OmegaConf.merge(tq_config, config)

client.initialize_storage_manager(manager_type="AsyncSimpleStorageManager", config=config)

return controller, storage_units, client

def demonstrate_random_sampler_with_replacement():

print("\n" + "=" * 80)

print("Example 1: Use Customized RandomSamplerWithReplacement in TransferQueue")

print("=" * 80)

print("\nSetup TransferQueue with RandomSamplerWithReplacement...")

sampler = RandomSamplerWithReplacement()

controller, storage_units, client = setup_transfer_queue_with_sampler(sampler)

# Add 5 samples

print("\n[Step 1] Adding 5 samples...")

data = TensorDict(

{

"input": torch.tensor([[i] for i in range(5)]),

batch_size=5,

)

client.put(data=data, partition_id="test")

print(" ✓ 5 samples added")

# Get batch 1 (should get 2 random samples)

print("\n[Step 2] Get batch 1 (2 samples)...")

meta1 = client.get_meta(data_fields=["input"], batch_size=2, partition_id="test", task_name="demo_task")

print(f" ✓ Got samples: {meta1.global_indexes}")

# Get batch 2 (should get 1 random sample with replacement - may have duplicate with previous batch!)

print("\n[Step 3] Get batch 2 (1 sample)...")

meta2 = client.get_meta(data_fields=["input"], batch_size=1, partition_id="test", task_name="demo_task")

print(f" ✓ Got samples: {meta2.global_indexes}")

# Get batch 3 (should get 2 random samples with replacement - may have duplicate with previous batches!)

print("\n[Step 4] Get batch 3 (2 samples)...")

meta3 = client.get_meta(data_fields=["input"], batch_size=2, partition_id="test", task_name="demo_task")

print(f" ✓ Got samples: {meta3.global_indexes}")

print("\n[Verification]")

print(" ✓ With replacement: Same sample can appear multiple times")

print(" ✓ Check: Are there duplicates in the batches?")

all_sampled = meta1.global_indexes + meta2.global_indexes + meta3.global_indexes

print(f" ✓ All sampled: {all_sampled}")

# Cleanup

client.clear_partition(partition_id="test")

client.close()

ray.shutdown()

def demonstrate_random_sampler_without_replacement():

print("\n" + "=" * 80)

print("Example 2: Use Customized RandomSamplerWithoutReplacement in TransferQueue")

print("=" * 80)

print("\nSetup TransferQueue with RandomSamplerWithoutReplacement...")

sampler = RandomSamplerWithoutReplacement()

controller, storage_units, client = setup_transfer_queue_with_sampler(sampler)

# Add 6 samples

print("\n[Step 1] Adding 6 samples...")

data = TensorDict(

{

"input": torch.tensor([[i] for i in range(6)]),

batch_size=6,

)

client.put(data=data, partition_id="test")

print(" ✓ 6 samples added")

# Get batch 1 (should get 3 random samples without replacement)

print("\n[Step 2] Get batch 1 (3 samples)...")

meta1 = client.get_meta(data_fields=["input"], batch_size=3, partition_id="test", task_name="demo_task")

print(f" ✓ Got samples: {meta1.global_indexes}")

# Get batch 2 (should randomly get 1 sample that are different from previous batch)

print("\n[Step 3] Get batch 2 (1 samples)...")

meta2 = client.get_meta(data_fields=["input"], batch_size=1, partition_id="test", task_name="demo_task")

print(f" ✓ Got samples: {meta2.global_indexes}")

# Get batch 3 (should randomly get 2 samples that are different from previous batch)

print("\n[Step 4] Get batch 3 (2 samples)...")

meta3 = client.get_meta(data_fields=["input"], batch_size=2, partition_id="test", task_name="demo_task")

print(f" ✓ Got samples: {meta3.global_indexes}")

print("\n[Verification]")

print(" ✓ Without replacement: Each sample consumed only once")

print(f" ✓ Batch 1: {meta1.global_indexes}")

print(f" ✓ Batch 2: {meta2.global_indexes}")

print(f" ✓ Batch 3: {meta3.global_indexes} (none left)")

# Cleanup

client.clear_partition(partition_id="test")

client.close()

ray.shutdown()

def demonstrate_priority_sampler():

print("\n" + "=" * 80)

print("Example 3: Use Customized PrioritySampler in TransferQueue")

print("=" * 80)

print("\nSetup TransferQueue with PrioritySampler...")

sampler = PrioritySampler()

controller, storage_units, client = setup_transfer_queue_with_sampler(sampler)

# Add 8 samples

print("\n[Step 1] Adding 8 samples...")

data = TensorDict(

{

"input": torch.tensor([[i] for i in range(8)]),

batch_size=8,

)

client.put(data=data, partition_id="test")

print(" ✓ 8 samples added")

time.sleep(1)

# Priority scores (higher = more important)

# Index 2, 7, 3 have highest priority

priority_scores = np.array([0.01, 0.01, 88, 999, 0.01, 0.01, 0.01, 10])

print("\n[Step 2] Get batch with priority (1 sample)...")

print(f"Priority scores: {priority_scores}")

# Get batch using priority sampling

meta1 = client.get_meta(

data_fields=["input"],

batch_size=1,

partition_id="test",

task_name="demo_task",

sampling_config={"priority_scores": priority_scores},

)

print(f" ✓ Got samples (high priority): {meta1.global_indexes}")

print(f" ✓ Priority of sampled: {[priority_scores[i] for i in meta1.global_indexes]}")

# Get another batch

print("\n[Step 3] Get another batch (2 samples)...")

meta2 = client.get_meta(

data_fields=["input"],

batch_size=2,

partition_id="test",

task_name="demo_task",

sampling_config={"priority_scores": priority_scores},

)

print(f" ✓ Got samples: {meta2.global_indexes}")

print(f" ✓ Priority of sampled: {[priority_scores[i] for i in meta2.global_indexes]}")

print("\n[Verification]")

print(" ✓ Priority sampling: Higher priority samples more likely to be chosen")

print(f" ✓ Batch 1 high-priority indices: {[i for i in meta1.global_indexes if priority_scores[i] >= 0.1]}")

print(f" ✓ Batch 2 high-priority indices: {[i for i in meta2.global_indexes if priority_scores[i] >= 0.1]}")

# Cleanup

client.clear_partition(partition_id="test")

client.close()

ray.shutdown()

def main():

print("=" * 80)

print(

textwrap.dedent(

"""

TransferQueue Tutorial 4: Custom Sampler Development

This script demonstrates how to develop custom samplers for TransferQueue.

Samplers control HOW data is consumed from the queue.

Core Interface:

- BaseSampler.sample(ready_indexes, batch_size, *args, **kwargs)

- Returns: (sampled_indexes, consumed_indexes)

- sampled_indexes has length = batch_size; consumed_indexes may be empty or have a different length

Key Concepts:

- ready_indexes: Samples ready for consumption (all fields produced & has not been consumed by the task)

- sampled_indexes: Which samples to return in this batch

- consumed_indexes: Which samples to mark as consumed (never returned to this task again)

"""

)

print("=" * 80)

try:

demonstrate_random_sampler_with_replacement()

demonstrate_random_sampler_without_replacement()

demonstrate_priority_sampler()

print("\n" + "=" * 80)

print("Tutorial Complete!")

print("=" * 80)

print("Key Takeaways:")

print("1. Samplers control HOW data is consumed from TransferQueue")

print("2. Implement BaseSampler.sample(ready_indexes, batch_size, *args, **kwargs)")

print("3. Return (sampled_indexes, consumed_indexes)")

print("4. Initialize controller with custom sampler: TransferQueueController.remote(sampler=YourSampler())")

print("5. Pass parameters via sampling_config in get_meta calls")

except Exception as e:

print(f"Error during tutorial: {e}")

import traceback

traceback.print_exc()

sys.exit(1)

if __name__ == "__main__":

main()

TransferQueue/tutorial/04_custom_sampler.py at dev · TransferQueue/TransferQueue · GitHub

Uh oh!

Expand file tree

Breadcrumbs

04_custom_sampler.py

Latest commit

History

Breadcrumbs

04_custom_sampler.py

File metadata and controls