Search results

Video MME Reproduce (Question about think mode usage)

Status: Open.

#39 In NVlabs/Long-RL;

· LimGeunTaekk opened on Dec 29, 2025

How should I run SFT using VILA codebase

Status: Open.

#38 In NVlabs/Long-RL;

· Yuanhong-Zheng opened on Dec 25, 2025

LongVideoReason Eval Dataset

Status: Open.

#37 In NVlabs/Long-RL;

· PREPONDERANCE opened on Dec 16, 2025

Where is the implementation of Stage1 parallel encoding (video frame uniform splitting)?

Status: Open.

#36 In NVlabs/Long-RL;

· HzZHoO opened on Dec 4, 2025

How to perform SFT on custom data

Status: Open.

#35 In NVlabs/Long-RL;

· sinbycos24 opened on Nov 20, 2025

Any plan to support Qwen3-VL dense model?

Status: Open.

#34 In NVlabs/Long-RL;

· RuixiangZhao opened on Oct 24, 2025

How to train with multi images?

Status: Open.

#33 In NVlabs/Long-RL;

· goodstudent9 opened on Sep 30, 2025

If this structure fit for NVILA-LITE-8B

Status: Open.

#32 In NVlabs/Long-RL;

· goodstudent9 opened on Sep 29, 2025

problem when run example/new_support/qwen2.5_omni_3b_grpo.sh

Status: Open.

#31 In NVlabs/Long-RL;

· wyclike opened on Sep 23, 2025

Why <code>log_probs.unsqueeze</code> in <code>compute_policy_loss</code>?

Status: Open.

#30 In NVlabs/Long-RL;

· Disapole-Xiao opened on Sep 18, 2025

The weights synchronizing of rollout

Status: Open.

#29 In NVlabs/Long-RL;

· coderwayne3025 opened on Sep 18, 2025

Can not run the Qwen-2.5-VL-7B model

Status: Open.

#28 In NVlabs/Long-RL;

· 64327069 opened on Sep 10, 2025