nanotron
/

doremi-llama-2.5b-reference

Model card Files Files and versions Community

doremi-llama-2.5b-reference / 70000 /optimizer /optimizer_config.json

neuralink's picture

neuralink HF staff

add the 70k checkpoint

4a65de7 10 months ago

124 Bytes

{"type": "OptimizerFromGradientAccumulator", "parallelism": {"tp_size": "8", "dp_size": "8", "pp_size": "1"}, "configs": {}}