Cosmos-RL

Supervised fine-tuning (SFT) of Cosmos Reason video QA models. The packaged default base model is hf_model://nvidia/Cosmos3-Nano. Pretrained weights are sourced from HuggingFace, not NGC. Gated HuggingFace models require HF_TOKEN. Some Cosmos-RL images cannot load the native Cosmos3 Omni checkpoint format directly; for those images, convert Cosmos3-Nano to a Qwen3-VL HF safetensors directory before train/evaluate and use that converted directory as the PTM path.

Uses FSDP-based parallelism with dp_shard_size for GPU count and dp_replicate_size for node count (not the standard num_gpus/num_nodes).

Requests for "Cosmos Reason 3", "Cosmos3 Nano Reasoner", or nvidia/Cosmos3-Nano are handled by this skill. There is no separate Cosmos3 model directory in the skill bank; route those requests here. Override the base HuggingFace model only when the user explicitly asks for a different model.

Deep detail lives in references; load the smallest one that matches the task: