Search: FSDP
14 result(s) on page 1
pytorch-fsdp
Fully sharded data-parallel training for large models.
quarantinedlinuxmacosDistributed Trainingoptional- Registry
- optional
- Category
- MLOps
- Version
- 1.0.0
- Author
- Orchestra Research
- License
- MIT
curl -s /v1/skills/community__axehub:optional__optional__pytorch-fsdpnemo-mbridge-perf-megatron-fsdp
Operational guide for enabling Megatron FSDP in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.
quarantinedClawHub- Registry
- ClawHub
- Category
- Training Ai· inferred
- Version
- 1.0.0
curl -s /v1/skills/community__axehub:ClawHub__ClawHub__nemo-mbridge-perf-megatron-fsdpnemo-mbridge-perf-megatron-fsdp
Operational guide for enabling Megatron FSDP in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.
quarantinedNVIDIA- Registry
- NVIDIA
- Category
- Training AI
- Version
- 1.0.0
curl -s /v1/skills/community__axehub:NVIDIA__NVIDIA__nemo-mbridge-perf-megatron-fsdpnemo-mbridge-perf-megatron-fsdp
Indexed by skills.sh from nvidia/skills
quarantinedskills.sh- Registry
- skills.sh
- Category
- Training Ai· inferred
- Version
- 1.0.0
- Author
- nvidia
curl -s /v1/skills/community__axehub:skills.sh__skills.sh__nemo-mbridge-perf-megatron-fsdppytorch-fsdp
Indexed by skills.sh from davila7/claude-code-templates
quarantinedskills.sh- Registry
- skills.sh
- Category
- Mlops· inferred
- Version
- 1.0.0
- Author
- davila7
curl -s /v1/skills/community__axehub:skills.sh__skills.sh__pytorch-fsdpaccelerate
Run PyTorch training across GPUs with minimal changes.
quarantinedlinuxmacoswindowsoptional- Registry
- optional
- Category
- MLOps
- Version
- 1.0.0
- Author
- Orchestra Research
- License
- MIT
curl -s /v1/skills/community__axehub:optional__optional__acceleratepytorch-lightning
Clean training loops with built-in distributed support.
quarantinedlinuxmacoswindowsoptional- Registry
- optional
- Category
- MLOps
- Version
- 1.0.0
- Author
- Orchestra Research
- License
- MIT
curl -s /v1/skills/community__axehub:optional__optional__pytorch-lightningpytorch-fsdp2
Indexed by skills.sh from orchestra-research/ai-research-skills
quarantinedskills.sh- Registry
- skills.sh
- Category
- Mlops· inferred
- Version
- 1.0.0
- Author
- orchestra-research
curl -s /v1/skills/community__axehub:skills.sh__skills.sh__pytorch-fsdp2nemo-automodel-distributed-training
Guide for selecting and configuring distributed training strategies in NeMo AutoModel, including FSDP2, Megatron FSDP, DDP, and parallelism settings.
quarantinedClawHub- Registry
- ClawHub
- Category
- Training Ai· inferred
- Version
- 1.0.0
curl -s /v1/skills/community__axehub:ClawHub__ClawHub__nemo-automodel-distributed-trainingnemo-mbridge-perf-moe-vlm-training
Practical guidance for training MoE VLMs in Megatron Bridge. Compares FSDP and 3D-parallel approaches, using rounded lessons from Qwen3-VL, Qwen3-Next, and other multimodal experiments.
quarantinedClawHub- Registry
- ClawHub
- Category
- Training Ai· inferred
- Version
- 1.0.0
curl -s /v1/skills/community__axehub:ClawHub__ClawHub__nemo-mbridge-perf-moe-vlm-trainingtao-finetune-cosmos-reason
Cosmos3-Nano video QA supervised fine-tuning with FSDP parallelism. Use when training or evaluating video question-answering models, fine-tuning Cosmos3-Nano or compatible Cosmos Reason models with SFT/LoRA, or working with Cosmos-RL. Trigger phrases include "fine-tune Cosmos"…
quarantinedClawHub- Registry
- ClawHub
- Version
- 1.0.0
curl -s /v1/skills/community__axehub:ClawHub__ClawHub__tao-finetune-cosmos-reasonnemo-automodel-distributed-training
Guide for selecting and configuring distributed training strategies in NeMo AutoModel, including FSDP2, Megatron FSDP, DDP, and parallelism settings.
quarantinedNVIDIA- Registry
- NVIDIA
- Category
- Training AI
- Version
- 1.0.0
curl -s /v1/skills/community__axehub:NVIDIA__NVIDIA__nemo-automodel-distributed-trainingnemo-mbridge-perf-moe-vlm-training
Practical guidance for training MoE VLMs in Megatron Bridge. Compares FSDP and 3D-parallel approaches, using rounded lessons from Qwen3-VL, Qwen3-Next, and other multimodal experiments.
quarantinedNVIDIA- Registry
- NVIDIA
- Category
- Training AI
- Version
- 1.0.0
curl -s /v1/skills/community__axehub:NVIDIA__NVIDIA__nemo-mbridge-perf-moe-vlm-trainingtao-finetune-cosmos-reason
Cosmos3-Nano video QA supervised fine-tuning with FSDP parallelism. Use when training or evaluating video question-answering models, fine-tuning Cosmos3-Nano or compatible Cosmos Reason models with SFT/LoRA, or working with Cosmos-RL. Trigger phrases include "fine-tune Cosmos"…
quarantinedvideoqacosmosNVIDIA