blender_3B.opt

{
  "init_opt": null,
  "show_advanced_args": false,
  "task": "interactive",
  "datatype": "test",
  "image_mode": "raw",
  "numthreads": 1,
  "hide_labels": false,
  "batchsize": 128,
  "dynamic_batching": null,
  "datapath": "ParlAI/data",
  "model": "transformer/generator",
  "model_file": "ParlAI/data/models/blender/blender_3B/model",
  "init_model": "/checkpoint/parlai/zoo/meena/20200319_meenav0data_tall_2.7B_adamoptimizer/20200319_13.3ppl_200kupdates/model",
  "dict_class": "parlai.core.dict:DictionaryAgent",
  "evaltask": null,
  "eval_batchsize": null,
  "display_examples": false,
  "num_epochs": -1,
  "max_train_time": 27647.999999999996,
  "validation_every_n_secs": -1,
  "save_every_n_secs": -1,
  "save_after_valid": true,
  "validation_every_n_epochs": 0.25,
  "validation_max_exs": -1,
  "short_final_eval": false,
  "validation_patience": 10,
  "validation_metric": "ppl",
  "validation_metric_mode": "min",
  "validation_cutoff": 1.0,
  "validation_share_agent": false,
  "metrics": "default",
  "aggregate_micro": false,
  "tensorboard_log": false,
  "dict_maxexs": -1,
  "dict_include_valid": false,
  "dict_include_test": false,
  "log_every_n_secs": 10.0,
  "image_size": 256,
  "image_cropsize": 224,
  "label_type": "response",
  "include_knowledge": true,
  "include_checked_sentence": true,
  "include_knowledge_separator": false,
  "num_topics": 5,
  "train_experiencer_only": false,
  "remove_political_convos": false,
  "embedding_size": 2560,
  "n_layers": 2,
  "ffn_size": 10240,
  "dropout": 0.1,
  "attention_dropout": 0.0,
  "relu_dropout": 0.0,
  "n_heads": 32,
  "learn_positional_embeddings": false,
  "embeddings_scale": true,
  "n_positions": 128,
  "n_segments": 0,
  "variant": "prelayernorm",
  "activation": "gelu",
  "output_scaling": 1.0,
  "share_word_embeddings": true,
  "n_encoder_layers": 2,
  "n_decoder_layers": 24,
  "model_parallel": true,
  "beam_size": 20,
  "beam_min_length": 20,
  "beam_context_block_ngram": 3,
  "beam_block_ngram": 3,
  "beam_length_penalty": 0.65,
  "skip_generation": false,
  "inference": "topk",
  "topk": 40,
  "topp": 0.9,
  "beam_delay": 30,
  "temperature": 1.0,
  "compute_tokenized_bleu": false,
  "embedding_type": "random",
  "embedding_projection": "random",
  "fp16": true,
  "fp16_impl": "mem_efficient",
  "force_fp16_tokens": true,
  "optimizer": "mem_eff_adam",
  "learningrate": 7e-06,
  "gradient_clip": 0.1,
  "adam_eps": 1e-08,
  "adafactor_eps": [
    1e-30,
    0.001
  ],
  "momentum": 0,
  "nesterov": true,
  "nus": [
    0.7
  ],
  "betas": [
    0.9,
    0.999
  ],
  "weight_decay": null,
  "rank_candidates": false,
  "truncate": 128,
  "text_truncate": 128,
  "label_truncate": 128,
  "history_size": -1,
  "person_tokens": false,
  "split_lines": false,
  "use_reply": "label",
  "add_p1_after_newln": false,
  "delimiter": "  ",
  "history_add_global_end_token": "end",
  "gpu": -1,
  "no_cuda": false,
  "dict_file": "ParlAI/data/models/blender/blender_3B/model.dict",
  "dict_initpath": null,
  "dict_language": "english",
  "dict_max_ngram_size": -1,
  "dict_minfreq": 0,
  "dict_maxtokens": -1,
  "dict_nulltoken": "__null__",
  "dict_starttoken": "__start__",
  "dict_endtoken": "__end__",
  "dict_unktoken": "__unk__",
  "dict_tokenizer": "bytelevelbpe",
  "dict_lower": false,
  "bpe_debug": false,
  "dict_textfields": "text,labels",
  "bpe_vocab": "ParlAI/data/models/blender/blender_3B/model.dict-vocab.json",
  "bpe_merge": "ParlAI/data/models/blender/blender_3B/model.dict-merges.txt",
  "bpe_add_prefix_space": true,
  "lr_scheduler": "reduceonplateau",
  "lr_scheduler_patience": 3,
  "lr_scheduler_decay": 0.5,
  "max_lr_steps": -1,
  "invsqrt_lr_decay_gamma": -1,
  "warmup_updates": 100,
  "warmup_rate": 0.0001,
  "update_freq": 2,
  "parlai_home": "ParlAI/",
  "starttime": "Mar31_06-04",
  "beam_block_full_context": false,
  "batchindex": 127,
  "dict_loaded": true
}