All issues

Search results

Support For Parallel Environment Steps

Status: Open.

#2301 In NovaSky-AI/SkyRL;

· amasin2111 opened on Sep 24, 2026

Add a dedicated documentation page for all SkyRL metrics

Status: Open.

#2300 In NovaSky-AI/SkyRL;

· SumanthRH opened on Sep 24, 2026

naive_dapo.compute_score: tiemout=True typo makes the \pi branch always raise (swallowed by bare except:), so correct numeric answers to \pi questions score 0.0

Status: Open.

#2289 In NovaSky-AI/SkyRL;

· shaurya416 opened on Sep 24, 2026

[tinker] Support newer Tinker SDKs: sampling fails with SDK 0.30.1

Status: Open.

#2287 In NovaSky-AI/SkyRL;

· luyuzhe111 opened on Sep 24, 2026

Support trainer-only and inference-only Tinker runtimes

Status: Open.

#2278 In NovaSky-AI/SkyRL;

· j316chuck opened on Sep 23, 2026

Support compatible alternate base weights for Tinker runtimes

Status: Open.

#2279 In NovaSky-AI/SkyRL;

· j316chuck opened on Sep 23, 2026

pad_training_input_batch pads TensorList fields with tensor[0].clone() — loss-masked pad rows carry a real row's payload (ragged-consumer crash; wrong pixel_values/image_grid_thw on pad rows)

Status: Open.

#2263 In NovaSky-AI/SkyRL;

· kunling-anyscale opened on Sep 22, 2026

fully_async (non-colocated): training actor emits all-NaN logits on every microbatch of the first step after the first NCCL weight sync; weights/optimizer/exports all finite

Status: Open.

#2247 In NovaSky-AI/SkyRL;

· kunling-anyscale opened on Sep 19, 2026

fully_async: non-colocated weight sync ignores clear_kv_cache_on_weight_sync — stale prefix-cache KV trained on after each sync

Status: Open.

#2246 In NovaSky-AI/SkyRL;

· kunling-anyscale opened on Sep 19, 2026

[bug][fsdp] FSDP2 manual offload/load corrupts DTensor unshard state under colocate_all=True, causing silent non-deterministic logit errors

Status: Open.

#2215 In NovaSky-AI/SkyRL;

· Aression opened on Sep 16, 2026

[feature] Honor max_tokens_per_microbatch in the loss-path forward (forward with loss_fn)

Status: Open.

#2211 In NovaSky-AI/SkyRL;

· avigyabb opened on Sep 15, 2026

[weight sync] Improve offload/backload semantics while training with LoRA with colocate_all=True

Status: Open.

#2206 In NovaSky-AI/SkyRL;

· SumanthRH opened on Sep 14, 2026