Search results

Publicly Releasing Cosmos-Predict2 rCM Weights

Status: Open.

#36 In NVlabs/rcm;

· OptimusLime opened on Jun 11, 2026

Using tf-dcm/tf-scm independently

Status: Open.

#35 In NVlabs/rcm;

· 834810269 opened on Jun 5, 2026

1-step better than 4-step & is JVP-based sCM really valuable？

Status: Open.

#34 In NVlabs/rcm;

· fy-zhang-ctrl opened on May 31, 2026

Question about Causal-rCM

Status: Open.

#33 In NVlabs/rcm;

· fy-zhang-ctrl opened on May 31, 2026

推理时固定的时间步选择和DMD训练时学生模拟去噪随机选择时间步之间的gap

Status: Open.

#32 In NVlabs/rcm;

· Men200 opened on Apr 15, 2026

DMD2 implementation

Status: Open.

#31 In NVlabs/rcm;

· Turlan opened on Apr 9, 2026

Question about discrete sCM loss vs standard CM loss performance

Status: Open.

#30 In NVlabs/rcm;

· SpursLipu opened on Mar 31, 2026

Clarification: training sampler (generate_samples_from_batch) vs inference sampler mismatch (TrigFlow vs RF) — is this intended?

Status: Open.

#27 In NVlabs/rcm;

· LeslieZhoa opened on Feb 4, 2026

Problems on Issues encountered during I2V rcm-training.

Status: Open.

#25 In NVlabs/rcm;

· ariyali opened on Feb 2, 2026

result of train wan-animate by rcm

Status: Open.

#24 In NVlabs/rcm;

· renrenzsbbb opened on Jan 30, 2026

the setting of tangent_warmup value

Status: Open.

#23 In NVlabs/rcm;

· renrenzsbbb opened on Jan 29, 2026

question about MLPProj(torch.nn.Module)

Status: Open.

#22 In NVlabs/rcm;

· renrenzsbbb opened on Jan 26, 2026