mt
Browse files- mt_template.json +80 -0
mt_template.json
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"exp_name":"mt_enfrde",
|
| 3 |
+
"exp_id":"demo",
|
| 4 |
+
"dump_path":"/content/dump_path",
|
| 5 |
+
"encoder_only":"False",
|
| 6 |
+
"data_path":"/content/processed",
|
| 7 |
+
"lgs":"de-en-fr",
|
| 8 |
+
"clm_steps":"",
|
| 9 |
+
"mlm_steps":"",
|
| 10 |
+
"mt_steps":"...",
|
| 11 |
+
"ae_steps":"",
|
| 12 |
+
"bt_steps":"",
|
| 13 |
+
"pc_steps":"",
|
| 14 |
+
"reload_emb":"",
|
| 15 |
+
"reload_model":"/content/dump_path/mlm_enfrde/demo/best-valid_mlm_ppl.pth,/content/dump_path/mlm_enfrde/demo/best-valid_mlm_ppl.pth",
|
| 16 |
+
"reload_checkpoint":"",
|
| 17 |
+
"epoch_size":"100",
|
| 18 |
+
"max_epoch":"1",
|
| 19 |
+
"stopping_criterion":"valid_en-fr_mt_bleu,10",
|
| 20 |
+
"validation_metrics":"valid_en-fr_mt_bleu",
|
| 21 |
+
"train_n_samples":"-1",
|
| 22 |
+
"valid_n_samples":"-1",
|
| 23 |
+
"test_n_samples":"-1",
|
| 24 |
+
"remove_long_sentences_train":"False",
|
| 25 |
+
"remove_long_sentences_valid":"False",
|
| 26 |
+
"remove_long_sentences_test":"False",
|
| 27 |
+
|
| 28 |
+
"eval_bleu":"True",
|
| 29 |
+
"eval_only":"False",
|
| 30 |
+
"max_len":"100",
|
| 31 |
+
"group_by_size":"True",
|
| 32 |
+
"batch_size":"32",
|
| 33 |
+
"save_periodic":"0",
|
| 34 |
+
"fp16":"False",
|
| 35 |
+
"amp":"-1",
|
| 36 |
+
"emb_dim":"1024",
|
| 37 |
+
"n_layers":"6",
|
| 38 |
+
"n_heads":"8",
|
| 39 |
+
"dropout":"0.1",
|
| 40 |
+
"attention_dropout":"0.1",
|
| 41 |
+
"gelu_activation":"True",
|
| 42 |
+
"share_inout_emb":"True",
|
| 43 |
+
"sinusoidal_embeddings":"False",
|
| 44 |
+
"use_lang_emb":"True",
|
| 45 |
+
"use_memory":"False",
|
| 46 |
+
"asm":"False",
|
| 47 |
+
"context_size":"0",
|
| 48 |
+
"word_pred":"0.15",
|
| 49 |
+
"word_mask_keep_rand":"0.8,0.1,0.1",
|
| 50 |
+
"word_shuffle":"3",
|
| 51 |
+
"word_dropout":"0.1",
|
| 52 |
+
"word_blank":"0.1",
|
| 53 |
+
"max_vocab":"-1",
|
| 54 |
+
"min_count":"0",
|
| 55 |
+
"lg_sampling_factor":"-1",
|
| 56 |
+
"bptt":"256",
|
| 57 |
+
"max_batch_size":"0",
|
| 58 |
+
"tokens_per_batch":"2000",
|
| 59 |
+
"split_data":"False",
|
| 60 |
+
"optimizer":"adam_inverse_sqrt,beta1=0.9,beta2=0.98,lr=0.0001",
|
| 61 |
+
"clip_grad_norm":"5",
|
| 62 |
+
"accumulate_gradients":"1",
|
| 63 |
+
"lambda_mlm":"1",
|
| 64 |
+
"lambda_clm":"1",
|
| 65 |
+
"lambda_pc":"1",
|
| 66 |
+
"lambda_ae":"0:1,100000:0.1,300000:0",
|
| 67 |
+
"lambda_mt":"1",
|
| 68 |
+
"lambda_bt":"1",
|
| 69 |
+
"beam_size":"1",
|
| 70 |
+
"length_penalty":"1",
|
| 71 |
+
"early_stopping":"False",
|
| 72 |
+
"debug_train":"False",
|
| 73 |
+
"debug_slurm":"False",
|
| 74 |
+
"local_rank":"-1",
|
| 75 |
+
"master_port":"-1",
|
| 76 |
+
"mem_enc_positions":"",
|
| 77 |
+
"mem_dec_positions":"",
|
| 78 |
+
"asm_cutoffs":"8000,20000",
|
| 79 |
+
"asm_div_value":"4"
|
| 80 |
+
}
|