Popular repositories Loading
-
pretrain-data
pretrain-data PublicPretraining data reconstruction scripts for Apertus
-
-
Megatron-LM
Megatron-LM PublicForked from NVIDIA/Megatron-LM
Ongoing research training transformer models at scale
-
Repositories
Showing 10 of 133 repositories
- Megatron-LM-MoE Public
-
-
- apertus-omni-tokenizer Public
- apertus-speculative-decoding Public
Reproducible performance study of speculative decoding for Apertus v1.5 on CSCS Clariden
- Megatron-Bridge Public Forked from NVIDIA-NeMo/Megatron-Bridge
Training library for Megatron-based models with bidirectional Hugging Face conversion capability
- hfconverter Public
- model-compatibility-suite Public
Model Compatibility Suite: quick, deterministic capability + conformance tests for a served (OpenAI-compatible) model endpoint
-
Most used topics
Loading…