Tunix - Browse /v0.1.7 at SourceForge.net

The interactive file manager requires Javascript. Please enable it or use sftp or scp.
You may still browse the files here.

Name	Modified	Size	InfoDownloads / Week
Parent folder
README.md	2026-06-09	28.0 kB	0
Tunix v0.1.7 source code.tar.gz	2026-06-09	28.9 MB	0
Tunix v0.1.7 source code.zip	2026-06-09	29.2 MB	0
Totals: 3 Items		58.2 MB	0

What's Changed

Instrument agentic loop with perf v2 & perfetto improvements by @copybara-service[bot] in https://github.com/google/tunix/pull/1159
add support for GCS endpoint for perf trace writing by @copybara-service[bot] in https://github.com/google/tunix/pull/1221
[Tunix] Expose many training hyperparameters via argparse. by @copybara-service[bot] in https://github.com/google/tunix/pull/1243
fix sglang warning by @copybara-service[bot] in https://github.com/google/tunix/pull/1242
fix: return type of _train_step() in PeftTrainer. by @copybara-service[bot] in https://github.com/google/tunix/pull/1241
remove extra <bos> from training prompt ids by @copybara-service[bot] in https://github.com/google/tunix/pull/1244
Uiuc vlm pr compressed fixed by @immuntasir in https://github.com/google/tunix/pull/511
[Tunix] Implement batch logging and add signal handlers for graceful shutdown for trajectory logger. by @copybara-service[bot] in https://github.com/google/tunix/pull/1245
measure batch solve_all & solve_none by @copybara-service[bot] in https://github.com/google/tunix/pull/1247
[Tunix] Update math eval script with vLLM sampler configuration to use vllm_jax backend and engine_kwargs. Adjust seeding for vLLm. by @copybara-service[bot] in https://github.com/google/tunix/pull/1246
[Tunix] Defaults to best mesh configurations. by @copybara-service[bot] in https://github.com/google/tunix/pull/1250
chore: start v0.1.7 development. by @copybara-service[bot] in https://github.com/google/tunix/pull/1248
use timeline snapshot for export to avoid race condition by @copybara-service[bot] in https://github.com/google/tunix/pull/1240
obtain old_logps in async rollout by @copybara-service[bot] in https://github.com/google/tunix/pull/1252
[Tunix] Add requirements.txt file with vllm and tpu-inference versions by @copybara-service[bot] in https://github.com/google/tunix/pull/1253
Add vllm rollout config for deepswe. by @copybara-service[bot] in https://github.com/google/tunix/pull/1222
Speed-up unscanning with jax unstack. by @NicoGrande in https://github.com/google/tunix/pull/1256
[Tunix] Remove eos token from user config. use default tokenizer.eos_id() directly. Besides for deepseek-distilled model we should have used <|end_of_sentence|> instead. by @copybara-service[bot] in https://github.com/google/tunix/pull/1267
fix copybara by @copybara-service[bot] in https://github.com/google/tunix/pull/1266
Support tunable parameter for submit train job in xpk by @copybara-service[bot] in https://github.com/google/tunix/pull/1269
[Tunix] Set max_logprobs to 1 in vLLM rollout. by @copybara-service[bot] in https://github.com/google/tunix/pull/1270
use background thread for writing trace by @copybara-service[bot] in https://github.com/google/tunix/pull/1249
fix oldprobs. Remove patching rollout config. Add validation check at agentic learner init. by @copybara-service[bot] in https://github.com/google/tunix/pull/1271
[Tunix] Fix aggregation functions for reward metrics. by @copybara-service[bot] in https://github.com/google/tunix/pull/1277
[Tunix][Oncall] Update expected output in vLLM sampler test. by @copybara-service[bot] in https://github.com/google/tunix/pull/1281
mixed precision for Qwen3 by @copybara-service[bot] in https://github.com/google/tunix/pull/1285
enable flash attn in Tunix model by @copybara-service[bot] in https://github.com/google/tunix/pull/1273
[Tunix] Update resharding to use latest pathwaysutils tests by @copybara-service[bot] in https://github.com/google/tunix/pull/1286
make lane assignment helper fn and fix naming by @copybara-service[bot] in https://github.com/google/tunix/pull/1272
Add more metrics for agentic rl. by @wang2yn84 in https://github.com/google/tunix/pull/1284
[JAX] Suppress type errors found by pytype after correcting definition of jax.typing.ArrayLike. by @copybara-service[bot] in https://github.com/google/tunix/pull/1254
[Tunix] Add jax.block_until_ready to test common utils. by @copybara-service[bot] in https://github.com/google/tunix/pull/1290
[Tunix] Clamp sequence importance ratio in agentic GRPO learner. by @copybara-service[bot] in https://github.com/google/tunix/pull/1296
[Tunix] Update expected output in vllm_sampler_test. by @copybara-service[bot] in https://github.com/google/tunix/pull/1300
pass role to device assignment to writer by @copybara-service[bot] in https://github.com/google/tunix/pull/1274
add flash attn to qwen3 model by @copybara-service[bot] in https://github.com/google/tunix/pull/1287
Disable progress bar if stderr is not TTY and not running in a notebook. by @copybara-service[bot] in https://github.com/google/tunix/pull/1292
remove step time metrics by @copybara-service[bot] in https://github.com/google/tunix/pull/1289
make async timeline failure log errors by @copybara-service[bot] in https://github.com/google/tunix/pull/1301
Role base track grouping by @copybara-service[bot] in https://github.com/google/tunix/pull/1261
[Tunix] Use self.config.return_logprobs to determine if logprobs should be returned. by @copybara-service[bot] in https://github.com/google/tunix/pull/1304
[Tunix] Add special handling for math answer grading. by @copybara-service[bot] in https://github.com/google/tunix/pull/1297
nit by @copybara-service[bot] in https://github.com/google/tunix/pull/1293
remove dummy reward fns for agentic learner. by @copybara-service[bot] in https://github.com/google/tunix/pull/1275
Add support for dynamically setting the number of steps for GRPO. by @niting in https://github.com/google/tunix/pull/1257
Add env time tracking by @copybara-service[bot] in https://github.com/google/tunix/pull/1302
Add megablox MoE support by @copybara-service[bot] in https://github.com/google/tunix/pull/1306
[Tunix] Hide internal helper functions for 1p only in reshard.py. by @copybara-service[bot] in https://github.com/google/tunix/pull/1291
[Tunix] Add backend_kwargs to MetricsLoggerOptions. by @copybara-service[bot] in https://github.com/google/tunix/pull/1305
add logging of example rewards with snipped output, controlled under debug=True flag from GRPOConfig by @copybara-service[bot] in https://github.com/google/tunix/pull/1298
Code update by @copybara-service[bot] in https://github.com/google/tunix/pull/1314
instrument environment interactions for perf metrics by @copybara-service[bot] in https://github.com/google/tunix/pull/1303
update metrics docs with new perf v2 api by @copybara-service[bot] in https://github.com/google/tunix/pull/1312
revert changes to reward_manager_test.py from b6e98aeb by @andytwigg in https://github.com/google/tunix/pull/1319
update news in readme by @copybara-service[bot] in https://github.com/google/tunix/pull/1320
fix qwen2 mixed precision & notebook by @copybara-service[bot] in https://github.com/google/tunix/pull/1315
[Tunix] Add support for loading models from Orbax checkpoints for evaluation of individual trained ckpts. by @copybara-service[bot] in https://github.com/google/tunix/pull/1318
[Tunix] Always log kl regardless if it's enabled. by @copybara-service[bot] in https://github.com/google/tunix/pull/1313
return logps in vanilla sampler by @copybara-service[bot] in https://github.com/google/tunix/pull/1328
Code update by @copybara-service[bot] in https://github.com/google/tunix/pull/1327
reward fn is only set at env reset. by @copybara-service[bot] in https://github.com/google/tunix/pull/1329
adding s-noghabi to github codeowners by @copybara-service[bot] in https://github.com/google/tunix/pull/1308
[Tunix] Update qwix dependency and TPU test image. by @copybara-service[bot] in https://github.com/google/tunix/pull/1145
enable sequence parallelism by @copybara-service[bot] in https://github.com/google/tunix/pull/1332
Add overlong filter to optionally filter overlong trajectory. by @copybara-service[bot] in https://github.com/google/tunix/pull/1283
Add reward time and env time. by @copybara-service[bot] in https://github.com/google/tunix/pull/1310
[Tunix] Update Pathways worker temporary storage and remote IDE port. by @copybara-service[bot] in https://github.com/google/tunix/pull/1343
fix deepscaler sort failure by @copybara-service[bot] in https://github.com/google/tunix/pull/1342
fix name of trace_dir in base_config by @copybara-service[bot] in https://github.com/google/tunix/pull/1346
add support for repeating kv head tensors during weight sync. by @NicoGrande in https://github.com/google/tunix/pull/1251
make cluster grouping names include multiple roles in perfetto by @copybara-service[bot] in https://github.com/google/tunix/pull/1347
Add num_batches to base_config.yaml. by @copybara-service[bot] in https://github.com/google/tunix/pull/1349
Adding force_compute_kl knob and fixing exception for kl computation by @copybara-service[bot] in https://github.com/google/tunix/pull/1344
[Tunix] Update expected output in vLLM sampler test. by @copybara-service[bot] in https://github.com/google/tunix/pull/1357
[Tunix] Add validation for vLLM server mode in AgenticRLLearner. by @copybara-service[bot] in https://github.com/google/tunix/pull/1354
[Tunix] Add configurable loss aggregation mode and KL loss mode to DeepScaler. by @copybara-service[bot] in https://github.com/google/tunix/pull/1355
instrument tunix main with perf v2 by @copybara-service[bot] in https://github.com/google/tunix/pull/1345
[Tunix]Adds a command-line argument for maximum prompt length and configures the metrics logger to log training hyperparameters to Weights & Biases (wandb). by @copybara-service[bot] in https://github.com/google/tunix/pull/1356
[Tunix] Add dual-clip on pg_loss in agentic GRPO learner and compute kl_loss before applying kl penalty. by @copybara-service[bot] in https://github.com/google/tunix/pull/1334
[Tunix] Refactor loss aggregation to handle empty sequences and improve test coverage. by @copybara-service[bot] in https://github.com/google/tunix/pull/1358
[Tunix] Update vLLM/tpu-inference commits. by @copybara-service[bot] in https://github.com/google/tunix/pull/1359
[Tunix] Fix the bug of partial write being caught by a read operation. by @copybara-service[bot] in https://github.com/google/tunix/pull/1363
make REMOTE_PW_PORT configurable by @copybara-service[bot] in https://github.com/google/tunix/pull/1360
enable flash attention option in cli. by @copybara-service[bot] in https://github.com/google/tunix/pull/1339
Remove "steps_per_sec" from tqdm train metrics. by @copybara-service[bot] in https://github.com/google/tunix/pull/1362
Tunix gemma4 by @copybara-service[bot] in https://github.com/google/tunix/pull/1365
gemma4 26b moe by @copybara-service[bot] in https://github.com/google/tunix/pull/1367
use State instead of GraphState by @copybara-service[bot] in https://github.com/google/tunix/pull/1369
update readme by @copybara-service[bot] in https://github.com/google/tunix/pull/1368
[Tunix] Add retry logic to Hugging Face model existence checks. by @copybara-service[bot] in https://github.com/google/tunix/pull/1374
mixed precision support in gemma4 by @copybara-service[bot] in https://github.com/google/tunix/pull/1372
[Tunix] Add Pathways Cloud support to grpo_main. by @copybara-service[bot] in https://github.com/google/tunix/pull/1371
Add huggingface dataset support. by @wang2yn84 in https://github.com/google/tunix/pull/1375
[Tunix] Remove rollout_vllm_swap_space_size_gb from Tunix. by @copybara-service[bot] in https://github.com/google/tunix/pull/1383
support flash attn in gemma4 by @copybara-service[bot] in https://github.com/google/tunix/pull/1376
add knobs for perf metrics in grpo_demo_llama3_qwen2.py by @copybara-service[bot] in https://github.com/google/tunix/pull/1348
move filtering degenerate groups to process results by @copybara-service[bot] in https://github.com/google/tunix/pull/1381
Avoid implicit conversion of jax arrays to dtypes. by @copybara-service[bot] in https://github.com/google/tunix/pull/1382
[Tunix] Add DeepScaleR CLI recipe by @wang2yn84 in https://github.com/google/tunix/pull/1373
Move RLOO advantage estimator from GRPOLearner to AgenticGRPOLearner. by @copybara-service[bot] in https://github.com/google/tunix/pull/1340
Remove the step_time parameter from the SFT TrainingHooks.on_train_step_end function. by @copybara-service[bot] in https://github.com/google/tunix/pull/1395
Convert deepswe to cli by @wang2yn84 in https://github.com/google/tunix/pull/1392
[Tunix] Add documentation for Trajectory Logging in Tunix. by @copybara-service[bot] in https://github.com/google/tunix/pull/1405
improve rl_cluster init corner case handling and test coverage by @copybara-service[bot] in https://github.com/google/tunix/pull/1311
Check if GCS path and raise error in v1 perf tracer by @copybara-service[bot] in https://github.com/google/tunix/pull/1389
Include vLLM and tpu inference to Tunix build. by @wang2yn84 in https://github.com/google/tunix/pull/1411
add multi-node setup doc by @copybara-service[bot] in https://github.com/google/tunix/pull/1412
Bug fix for trajectory engine by @copybara-service[bot] in https://github.com/google/tunix/pull/1394
Fix broken base_config for RL training. by @niting in https://github.com/google/tunix/pull/1413
Add DeepSWE evaluation script with ActionGuard. by @copybara-service[bot] in https://github.com/google/tunix/pull/1406
mock rollout engine by @copybara-service[bot] in https://github.com/google/tunix/pull/1391
Allow various length distributions in mock rollout by @copybara-service[bot] in https://github.com/google/tunix/pull/1398
update docs with mock rollout engine by @copybara-service[bot] in https://github.com/google/tunix/pull/1400
[tunix/sft] Avoid eager re-sharding of globally distributed arrays by @copybara-service[bot] in https://github.com/google/tunix/pull/1325
Fix agentic cli dataloading by @wang2yn84 in https://github.com/google/tunix/pull/1416
Fix multiple dataset issue. by @wang2yn84 in https://github.com/google/tunix/pull/1419
check if limit reached more proactivately by @copybara-service[bot] in https://github.com/google/tunix/pull/1388
Internal by @copybara-service[bot] in https://github.com/google/tunix/pull/1390
add test for export failing by @copybara-service[bot] in https://github.com/google/tunix/pull/1421
[Tunix] Bump up the vllm/tpu-inference commit by @copybara-service[bot] in https://github.com/google/tunix/pull/1424
Adding support for fused moe kernels in sampler. by @NicoGrande in https://github.com/google/tunix/pull/1385
Save timelines per step by @copybara-service[bot] in https://github.com/google/tunix/pull/1423
use current and commited step in perf tracer by @copybara-service[bot] in https://github.com/google/tunix/pull/1425
[Tunix] Update Dockerfile dependencies and installation order. by @copybara-service[bot] in https://github.com/google/tunix/pull/1428
add timeline sequentializing to perf tracer by @copybara-service[bot] in https://github.com/google/tunix/pull/1422
Update training dataset to R2E-Gym-Subset. by @copybara-service[bot] in https://github.com/google/tunix/pull/1426
Prevent adding environment tokens for terminal steps in trajectory collection. by @copybara-service[bot] in https://github.com/google/tunix/pull/1427
Fix deleting destination buffers during reshard. by @NicoGrande in https://github.com/google/tunix/pull/1430
Add sequence packing by @copybara-service[bot] in https://github.com/google/tunix/pull/1337
[Tunix] Update logging for Pathways checkpointing based on persistence API usage. by @copybara-service[bot] in https://github.com/google/tunix/pull/1434
Disable sglang_jax with LoRA test in nightly regression by @sizhit2 in https://github.com/google/tunix/pull/1435
Fix env logging metrics tab by @sizhit2 in https://github.com/google/tunix/pull/1437
add model registry by @copybara-service[bot] in https://github.com/google/tunix/pull/1446
algo fix by @copybara-service[bot] in https://github.com/google/tunix/pull/1450
fix error message for max steps by @copybara-service[bot] in https://github.com/google/tunix/pull/1453
set vllm logprobs_mode by @copybara-service[bot] in https://github.com/google/tunix/pull/1451
Fix RL cluster mesh logic for null objects + Update model name in Gemma2 config for SFT and RL and add appropriate data_source. by @copybara-service[bot] in https://github.com/google/tunix/pull/1444
update nightly regression test by @copybara-service[bot] in https://github.com/google/tunix/pull/1455
add gemma 4 to tests by @copybara-service[bot] in https://github.com/google/tunix/pull/1447
Integrate MaxText Model Source in AutoModel by @sizhit2 in https://github.com/google/tunix/pull/1442
minor speedup by @copybara-service[bot] in https://github.com/google/tunix/pull/1460
Disable failed vLLM test in TPU nightly regression. by @copybara-service[bot] in https://github.com/google/tunix/pull/1459
[Tunix] Fix the nightly regression test path. by @copybara-service[bot] in https://github.com/google/tunix/pull/1463
[Tunix] Add the repository root to sys.path, so from tunix.cli import grpo_main resolves against the local tunix package in the workspace by @copybara-service[bot] in https://github.com/google/tunix/pull/1464
[Tunix] Remove pytree=False from Llama3 model. by @copybara-service[bot] in https://github.com/google/tunix/pull/1466
populate deepswe train branch to main by @copybara-service[bot] in https://github.com/google/tunix/pull/1420
Remove @nnx.jit from Gemma4 model call. by @copybara-service[bot] in https://github.com/google/tunix/pull/1468
Adding support for TP for fused MoE kernel. by @NicoGrande in https://github.com/google/tunix/pull/1448
Fix Arbitrary Code Execution via eval() in Reward Calculation. by @copybara-service[bot] in https://github.com/google/tunix/pull/1471
fix qwen parser thinking mode by @copybara-service[bot] in https://github.com/google/tunix/pull/1475
[Tunix] Add the missing train_fraction to cli scripts. by @copybara-service[bot] in https://github.com/google/tunix/pull/1477
[Tunix] Fix the lora filter for llama and qwen. by @copybara-service[bot] in https://github.com/google/tunix/pull/1478
Update nightly run env by @copybara-service[bot] in https://github.com/google/tunix/pull/1470
bug fix for agentic reward calculation. by @copybara-service[bot] in https://github.com/google/tunix/pull/1472
Fixed a test failure. by @copybara-service[bot] in https://github.com/google/tunix/pull/1473
Fix Command Injection in TPU single host script. by @copybara-service[bot] in https://github.com/google/tunix/pull/1480
[Tunix] Unpin jax and flax versions for sglnag tests in CI. by @copybara-service[bot] in https://github.com/google/tunix/pull/1483
Disable more cli tests in nightly regression and add TODO by @copybara-service[bot] in https://github.com/google/tunix/pull/1492
add control token sanitization to prevent prompt injection. by @copybara-service[bot] in https://github.com/google/tunix/pull/1489
set download path for all cli scripts by @copybara-service[bot] in https://github.com/google/tunix/pull/1491
Remove context_ratio from agentic GRPO config. by @copybara-service[bot] in https://github.com/google/tunix/pull/1495
Dynamic batch should be write back to config if it's inferred. by @wang2yn84 in https://github.com/google/tunix/pull/1479
Fix flash attention issue on decode path. by @wang2yn84 in https://github.com/google/tunix/pull/1481
Add weights mappings and sync from Tunix to VLLM-tpu-jax for Gemma4 models. by @copybara-service[bot] in https://github.com/google/tunix/pull/1486
Add "seq-mean-token-sum" loss aggregation mode. by @copybara-service[bot] in https://github.com/google/tunix/pull/1476
Fix mesh in cli scripts by @copybara-service[bot] in https://github.com/google/tunix/pull/1485
[Tunix] Remove trailing whitespace in cpu-tests.yml. by @copybara-service[bot] in https://github.com/google/tunix/pull/1500
Add degenerate_group_masking option to DeepSWE training script. by @copybara-service[bot] in https://github.com/google/tunix/pull/1488
make BaseTool immutable. by @copybara-service[bot] in https://github.com/google/tunix/pull/1497
Relax the constraint on batch dim for computing ref_logps by @copybara-service[bot] in https://github.com/google/tunix/pull/1431
Add sharding to the embedder parameters in Gemma4 model. by @copybara-service[bot] in https://github.com/google/tunix/pull/1498
Allow configurable micro-batching for compute_logps in agentic RL. by @copybara-service[bot] in https://github.com/google/tunix/pull/1436
Bump qwix version to 0.1.6 by @copybara-service[bot] in https://github.com/google/tunix/pull/1504
Add option to recompute old policy logprobs on the trainer. by @copybara-service[bot] in https://github.com/google/tunix/pull/1487
Specify shell: bash for TPU nightly RL script by @copybara-service[bot] in https://github.com/google/tunix/pull/1507
yaml based configs for llama cli scripts by @copybara-service[bot] in https://github.com/google/tunix/pull/1452
[Tunix] Update Dockerfile to fix qwix version conflict. by @copybara-service[bot] in https://github.com/google/tunix/pull/1509
Standardize Tunix model naming to align with HuggingFace conventions. by @copybara-service[bot] in https://github.com/google/tunix/pull/1433
yaml based configs for gemma3 cli scripts by @copybara-service[bot] in https://github.com/google/tunix/pull/1508
Stabilize Tunix SFT pipelines: resolve NaN divergence and persistent state corruption by @copybara-service[bot] in https://github.com/google/tunix/pull/1511
Fix non agentic colocate mode in cli. by @wang2yn84 in https://github.com/google/tunix/pull/1457
Add chunked resharding and target buffer deletion options for tunix models' weight sync. by @copybara-service[bot] in https://github.com/google/tunix/pull/1499
[Tunix] minor bug fix -- respect completion_mask when _compute_ref_per_token_logps. by @copybara-service[bot] in https://github.com/google/tunix/pull/1501
add orpo_trainer test to build file by @copybara-service[bot] in https://github.com/google/tunix/pull/1515
Add run_latest CI job for Tunix nightly regression by @copybara-service[bot] in https://github.com/google/tunix/pull/1510
Remove stale context_ratio from GRPO configurations. by @copybara-service[bot] in https://github.com/google/tunix/pull/1519
fix PEFT trainer early exit by @copybara-service[bot] in https://github.com/google/tunix/pull/1418
fix naming in scripts by @copybara-service[bot] in https://github.com/google/tunix/pull/1521
Fix the nightly regression to report error status correctly. by @wang2yn84 in https://github.com/google/tunix/pull/1522
tighten atol rtol on common_test by @copybara-service[bot] in https://github.com/google/tunix/pull/1524
Make it easier to see individual test results in TPU nightly by @copybara-service[bot] in https://github.com/google/tunix/pull/1525
Support remat for gemma4 by @copybara-service[bot] in https://github.com/google/tunix/pull/1531
[Tunix] Replace HTML-based image centering with standard g3doc Markdown image syntax. by @copybara-service[bot] in https://github.com/google/tunix/pull/1503
[Tunix] replace info logging to debug logging on state. by @copybara-service[bot] in https://github.com/google/tunix/pull/1527
Adjust top_p sampling to compute top-p of top-k logits, rather than top-k of top-p logits, matching vllm implementation by @copybara-service[bot] in https://github.com/google/tunix/pull/1528
support gemma4 in automodel by @copybara-service[bot] in https://github.com/google/tunix/pull/1533
Agentic GRPO: TIS correction, eval dedup, flash-attn segment_ids by @colincai-mc in https://github.com/google/tunix/pull/1523
[Tunix] Update TPU nightly regression workflow to build vLLM/tpu-inference from source. by @copybara-service[bot] in https://github.com/google/tunix/pull/1532
Gemma4 sampling optimizations. by @copybara-service[bot] in https://github.com/google/tunix/pull/1415
[Tunix] Fix weight mapping for gemma4 E2B model. by @copybara-service[bot] in https://github.com/google/tunix/pull/1540
Simplify logps compute by @copybara-service[bot] in https://github.com/google/tunix/pull/1548
[Tunix]Ensure anchor policy state is on device before resharding as pathways reshard doesn't support resharding between different memory kinds. by @copybara-service[bot] in https://github.com/google/tunix/pull/1549
Enhance performance and avoid OOM for data parallelism over a large number of devices O(100) by @copybara-service[bot] in https://github.com/google/tunix/pull/1550
[Tunix] minor fix on anchor trainer state. by @copybara-service[bot] in https://github.com/google/tunix/pull/1557
[Tunix] Default degenerate_group_masking to false. by @copybara-service[bot] in https://github.com/google/tunix/pull/1559
Add orbax-checkpoint to tunix dependencies. by @copybara-service[bot] in https://github.com/google/tunix/pull/1462
Introduce checkpoint_options to manage options for V1 Orbax Checkpointing. by @copybara-service[bot] in https://github.com/google/tunix/pull/1408
Migrate Tunix's usage of Orbax V0 Checkpoint Manager to V1 Checkpointer. by @copybara-service[bot] in https://github.com/google/tunix/pull/1409
Update Checkpointing Options construction through CLI configuration to move from Orbax V0 CheckpointManagerOptions to Tunix's CheckpointOptions. by @copybara-service[bot] in https://github.com/google/tunix/pull/1410
[Tunix] Ensure TITO for prompt tokens. Now the trajectory prompt tokens is updated to what's used for rollout sampler. by @copybara-service[bot] in https://github.com/google/tunix/pull/1561
Fix GSM8K reward function logic and string metrics logging. by @copybara-service[bot] in https://github.com/google/tunix/pull/1558
Add Gemma 4 E2B config and support for local HuggingFace tokenizers. by @copybara-service[bot] in https://github.com/google/tunix/pull/1562
Add optional output_clamp_value to compute_kl_divergence by @py4 in https://github.com/google/tunix/pull/1560
Update Checkpointing Options construction through CLI configuration to move from Orbax V0 CheckpointManagerOptions to Tunix's CheckpointOptions. by @copybara-service[bot] in https://github.com/google/tunix/pull/1563
Migrate Tunix's usage of Orbax V0 Checkpoint Manager to V1 Checkpointer. by @copybara-service[bot] in https://github.com/google/tunix/pull/1564
Fixes the tfds prompt not sanitizated issue for agentic path. by @wang2yn84 in https://github.com/google/tunix/pull/1556
Topology aware mesh creation by @wang2yn84 in https://github.com/google/tunix/pull/1465
Fix Tunix package test by @copybara-service[bot] in https://github.com/google/tunix/pull/1568
Fix run_prod CI: cap safetensors<0.8 to fix .ctypes breakage by @copybara-service[bot] in https://github.com/google/tunix/pull/1572
Add Gemma 4 Orbax checkpoint loading support to Tunix. by @copybara-service[bot] in https://github.com/google/tunix/pull/1553
pin safetensors in project toml by @copybara-service[bot] in https://github.com/google/tunix/pull/1573
Undo transformers<=4.57.1 pin by @ringohoffman in https://github.com/google/tunix/pull/1566