{ "valid_metrics": [ "BLEU" ], "n_sample": 0, "tgt_vocab_size": 20000, "vocab_size_multiple": 8, "save_data": "data", "transforms": [ "sentencepiece", "filtertoolong" ], "overwrite": true, "src_vocab_size": 20000, "tensorboard_log_dir_dated": "tensorboard/Sep-01_20-43-08", "share_vocab": false, "tgt_vocab": "th.eole.vocab", "report_every": 100, "tensorboard_log_dir": "tensorboard", "src_vocab": "en.eole.vocab", "tensorboard": true, "seed": 1234, "training": { "accum_steps": [ 0 ], "model_path": "quickmt-en-th-eole-model", "max_grad_norm": 0.0, "dropout_steps": [ 0 ], "learning_rate": 2.0, "decay_method": "noam", "accum_count": [ 10 ], "batch_type": "tokens", "attention_dropout": [ 0.1 ], "gpu_ranks": [ 0 ], "compute_dtype": "torch.float16", "prefetch_factor": 32, "bucket_size": 128000, "num_workers": 0, "batch_size_multiple": 8, "optim": "adamw", "train_steps": 100000, "average_decay": 0.0001, "normalization": "tokens", "param_init_method": "xavier_uniform", "label_smoothing": 0.1, "batch_size": 8000, "warmup_steps": 4000, "valid_batch_size": 4096, "valid_steps": 5000, "adam_beta2": 0.998, "world_size": 1, "dropout": [ 0.1 ], "save_checkpoint_steps": 5000, "keep_checkpoint": 4 }, "transforms_configs": { "filtertoolong": { "src_seq_length": 256, "tgt_seq_length": 256 }, "sentencepiece": { "src_subword_model": "${MODEL_PATH}/en.spm.model", "tgt_subword_model": "${MODEL_PATH}/th.spm.model" } }, "data": { "corpus_1": { "path_src": "train.en", "path_align": null, "path_tgt": "train.th", "transforms": [ "sentencepiece", "filtertoolong" ] }, "valid": { "path_src": "valid.en", "path_align": null, "path_tgt": "valid.th", "transforms": [ "sentencepiece", "filtertoolong" ] } }, "model": { "hidden_size": 1024, "share_embeddings": false, "position_encoding_type": "SinusoidalInterleaved", "share_decoder_embeddings": false, "transformer_ff": 4096, "heads": 8, "architecture": "transformer", "embeddings": { "src_word_vec_size": 1024, "word_vec_size": 1024, "position_encoding_type": "SinusoidalInterleaved", "tgt_word_vec_size": 1024 }, "decoder": { "tgt_word_vec_size": 1024, "hidden_size": 1024, "decoder_type": "transformer", "n_positions": null, "position_encoding_type": "SinusoidalInterleaved", "transformer_ff": 4096, "heads": 8, "layers": 2 }, "encoder": { "hidden_size": 1024, "encoder_type": "transformer", "src_word_vec_size": 1024, "n_positions": null, "position_encoding_type": "SinusoidalInterleaved", "transformer_ff": 4096, "heads": 8, "layers": 8 } } }