Download sft_config.json from kerzgrr/Monostich-2: direct link, hf CLI and curl.
- Browser
- Download file 1.12 kB
-
https://huggingface.co/kerzgrr/Monostich-2/resolve/main/sft_config.json
- Command line
-
hf download hf://kerzgrr/Monostich-2/sft_config.json
-
curl -L -o sft_config.json https://huggingface.co/kerzgrr/Monostich-2/resolve/main/sft_config.json
1.12 kB
| { | |
| "adam_beta1": 0.9, | |
| "adam_beta2": 0.95, | |
| "adam_epsilon": 1e-08, | |
| "cpu_prefetch_factor": 2, | |
| "cpu_workers": 2, | |
| "dataset_manifest": "data/tokenized/smoltalk-hermes-sft/manifest.json", | |
| "ema_inv_gamma": 1.0, | |
| "ema_max_decay": 0.9999, | |
| "ema_power": 0.75, | |
| "enable_gradient_checkpointing": true, | |
| "full_shuffle": true, | |
| "gradient_accumulation_steps": 16, | |
| "gradient_clip_norm": 1.0, | |
| "initial_run_dir": "runs/tiny-gdn-150m-smoltalk-sft", | |
| "initial_weights": "ema", | |
| "learning_rate": 0.0001, | |
| "log_every_optimizer_steps": 1, | |
| "logit_z_loss_coefficient": 0.0001, | |
| "maximum_checkpoints": 5, | |
| "maximum_optimizer_steps": null, | |
| "micro_batch_size": 1, | |
| "minimum_learning_rate_ratio": 0.1, | |
| "muon_learning_rate": 0.01, | |
| "muon_momentum": 0.95, | |
| "optimizer": "adamw", | |
| "output_dir": "runs/tiny-gdn-150m-smoltalk-hermes-sft", | |
| "require_complete_pretraining": false, | |
| "require_complete_sft_source": false, | |
| "save_every_optimizer_steps": 100, | |
| "seed": 20260720, | |
| "sequence_length": 8192, | |
| "shuffle_block_sequences": 512, | |
| "validation_batches": 16, | |
| "warmup_ratio": 0.01, | |
| "weight_decay": 0.0 | |
| } | |