{ "tensorboard": true, "seed": 1234, "tgt_vocab": "id.eole.vocab", "overwrite": true, "tensorboard_log_dir": "tensorboard", "n_sample": 0, "vocab_size_multiple": 8, "report_every": 100, "transforms": [ "sentencepiece", "filtertoolong" ], "src_vocab_size": 20000, "valid_metrics": [ "BLEU" ], "share_vocab": false, "tgt_vocab_size": 20000, "save_data": "data", "tensorboard_log_dir_dated": "tensorboard/May-28_20-39-34", "src_vocab": "en.eole.vocab", "training": { "normalization": "tokens", "attention_dropout": [ 0.1 ], "gpu_ranks": [ 0 ], "max_grad_norm": 0.0, "decay_method": "noam", "accum_steps": [ 0 ], "keep_checkpoint": 4, "batch_size_multiple": 8, "prefetch_factor": 32, "valid_batch_size": 4096, "learning_rate": 2.0, "batch_size": 8000, "warmup_steps": 4000, "compute_dtype": "torch.float16", "valid_steps": 5000, "bucket_size": 128000, "average_decay": 0.0001, "world_size": 1, "dropout": [ 0.1 ], "save_checkpoint_steps": 5000, "dropout_steps": [ 0 ], "train_steps": 100000, "optim": "adamw", "adam_beta2": 0.998, "num_workers": 0, "model_path": "quickmt-en-id-eole-model", "batch_type": "tokens", "accum_count": [ 10 ], "param_init_method": "xavier_uniform", "label_smoothing": 0.1 }, "transforms_configs": { "sentencepiece": { "src_subword_model": "${MODEL_PATH}/en.spm.model", "tgt_subword_model": "${MODEL_PATH}/id.spm.model" }, "filtertoolong": { "tgt_seq_length": 256, "src_seq_length": 256 } }, "data": { "corpus_1": { "transforms": [ "sentencepiece", "filtertoolong" ], "path_src": "train.en", "path_tgt": "train.id", "path_align": null }, "valid": { "transforms": [ "sentencepiece", "filtertoolong" ], "path_src": "dev.en", "path_tgt": "dev.id", "path_align": null } }, "model": { "share_embeddings": false, "transformer_ff": 4096, "position_encoding_type": "SinusoidalInterleaved", "share_decoder_embeddings": false, "hidden_size": 1024, "heads": 8, "architecture": "transformer", "embeddings": { "tgt_word_vec_size": 1024, "src_word_vec_size": 1024, "position_encoding_type": "SinusoidalInterleaved", "word_vec_size": 1024 }, "decoder": { "n_positions": null, "transformer_ff": 4096, "position_encoding_type": "SinusoidalInterleaved", "tgt_word_vec_size": 1024, "decoder_type": "transformer", "hidden_size": 1024, "heads": 8, "layers": 2 }, "encoder": { "src_word_vec_size": 1024, "n_positions": null, "transformer_ff": 4096, "position_encoding_type": "SinusoidalInterleaved", "encoder_type": "transformer", "hidden_size": 1024, "heads": 8, "layers": 8 } } }