| { | |
| "ngpus": 32, | |
| "tokens": 50257, | |
| "gpt_dir": "assets/gpt2-large", | |
| "outdir": "../output", | |
| "training": { | |
| "batch_size": 512, | |
| "accum": 1, | |
| "n_iters": 1000001, | |
| "snapshot_freq": 50000, | |
| "log_freq": 50, | |
| "eval_freq": 100, | |
| "snapshot_freq_for_preemption": 10000, | |
| "weight": "standard", | |
| "snapshot_sampling": false, | |
| "ema": 0.9999, | |
| "loss_type": "lambda_DCE" | |
| }, | |
| "data": { | |
| "train": "openwebtext", | |
| "valid": "wikitext103", | |
| "cache_dir": "data" | |
| }, | |
| "noise": { | |
| "type": "loglinear", | |
| "sigma_min": 0.0001, | |
| "sigma_max": 20 | |
| }, | |
| "sampling": { | |
| "predictor": "euler", | |
| "steps": 1024 | |
| }, | |
| "eval": { | |
| "batch_size": 512, | |
| "perplexity": true, | |
| "perplexity_batch_size": 16 | |
| }, | |
| "optim": { | |
| "weight_decay": 0.03, | |
| "optimizer": "AdamW", | |
| "lr": 0.0003, | |
| "beta1": 0.9, | |
| "beta2": 0.999, | |
| "eps": 1e-08, | |
| "warmup": 2500, | |
| "grad_clip": 1.0 | |
| }, | |
| "model": { | |
| "name": "small_wotsm", | |
| "type": "ddit_wot", | |
| "hidden_size": 768, | |
| "cond_dim": 128, | |
| "length": 1024, | |
| "n_blocks": 12, | |
| "n_heads": 12, | |
| "dropout": 0.02, | |
| "use_checkpoint": false, | |
| "dtype": "float16" | |
| } | |
| } |