mt5-base-koquad-ae / trainer_config.json
asahi417's picture
model update
d690cec
raw
history blame
352 Bytes
{"dataset_path": "lmqg/qg_koquad", "dataset_name": "default", "input_types": ["paragraph_sentence"], "output_types": ["answer"], "prefix_types": null, "model": "google/mt5-base", "max_length": 512, "max_length_output": 32, "epoch": 5, "batch": 8, "lr": 0.0005, "fp16": false, "random_seed": 1, "gradient_accumulation_steps": 8, "label_smoothing": 0.15}