| { |
| "variant": "nemotron_climb_proxy_model_350m", |
| "container_sif": "/home/pj26000041/ku60000248/work/climbmix-ja/artifacts/containers/nemo_26.02_nemotron_3_super.sif", |
| "pretrained_model_path": "/home/pj26000041/ku60000248/work/climbmix-ja/artifacts/nemotron-climb-proxy-models/nemotron_climb_proxy_model_350m_megatron_converted_te_compat", |
| "tokenizer_model": "/home/pj26000041/ku60000248/work/climbmix-ja/artifacts/tokenizers/llama2/tokenizer.model", |
| "indexed_text_prefix": "/home/pj26000041/ku60000248/work/climbmix-ja/artifacts/climblab-ja/proxy_megatron_llama2_balanced/indexed/climblab_ja_balanced_text_document", |
| "mixture_script": "/home/pj26000041/ku60000248/work/climbmix-ja/artifacts/climbsearch/climblab_ja_initial64_qwen_prior_temp0p5_cap100b/n21.sh", |
| "train_data_args_file": "/home/pj26000041/ku60000248/work/climbmix-ja/outputs/climbsearch_initial64_350m_tecompat_lr75e4_fixed6500_decay6500_20260530/n21/train_data_path.args", |
| "train_data_manifest": "/home/pj26000041/ku60000248/work/climbmix-ja/outputs/climbsearch_initial64_350m_tecompat_lr75e4_fixed6500_decay6500_20260530/n21/train_data_path.tsv", |
| "output_root": "/home/pj26000041/ku60000248/work/climbmix-ja/outputs/climbsearch_initial64_350m_tecompat_lr75e4_fixed6500_decay6500_20260530/n21", |
| "seq_len": 1024, |
| "train_iters": 6500, |
| "train_exit_min": 0, |
| "micro_batch_size": 76, |
| "global_batch_size": 304, |
| "lr": "7.5e-4", |
| "min_lr": "7.5e-5", |
| "lr_warmup_fraction": "0.01", |
| "lr_decay_iters": 6500, |
| "lr_decay_style": "cosine", |
| "nproc_per_node": 4, |
| "master_port": 29500, |
| "transformer_impl": "transformer_engine", |
| "use_flash_attn": "1", |
| "pytorch_cuda_alloc_conf": "expandable_segments:True", |
| "save_checkpoint": "1", |
| "save_interval": 50000000, |
| "extra_megatron_args": "--use-distributed-optimizer --overlap-grad-reduce --overlap-param-gather" |
| } |
|
|