Hugging Face's logo

Spaces:
rynmurdock
/
Babel
Runtime error

App Files Community
Babel
File size: 4,924 Bytes 
c5ca37a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
 
export PYTHONPATH="${PYTHONPATH}:/workspace/code"

export TRAIN_FILE=../data/datasets/dailydialog_data/train.txt
export TEST_FILE=../data/datasets/dailydialog_data/test.txt
export GENERATED_TEXT_FILE=../output/local_dialog_dataloader/eval_text_generation_results.txt

export GPU_ID=0,1

CUDA_VISIBLE_DEVICES=$GPU_ID python examples/big_ae/run_spacefusion_pretraining.py \
    --dataset dailydialog \
    --output_dir=../output/local_dialog_dataloader \
    --encoder_model_type=bert \
    --encoder_model_name_or_path=bert-base-cased \
    --decoder_model_type=gpt2 \
    --decoder_model_name_or_path=gpt2 \
    --train_data_file=$TRAIN_FILE \
    --do_generation \
    --do_train \
    --do_eval \
    --beta 2.0 \
    --ratio_zero .5 \
    --ratio_increase 0.25 \
    --eval_data_file=$TEST_FILE \
    --num_train_epochs 1.0 \
    --save_steps 2000 \
    --logging_steps 100 \
    --overwrite_output_dir \
    --per_gpu_train_batch_size 4 \
    --block_size 512 \
    --freeze_bert \
    --per_gpu_eval_batch_size 1 \
    --total_sents -1 \
    --sents_per_cxt 10 \
    --num_frozen_bert_layer 10 \
    --num_s2s_bert_layer 2 \
    --eval_generated_text_file $GENERATED_TEXT_FILE\
    --checkpoint_dir ../output/philly_rr3scl_g8_vae_wikipedia_pretraining_beta_schedule_beta1.0_d1.0_ro0.5_ra0.25 \
    --gloabl_step_eval 760000  \
    --use_pretrained_model \
    --use_pretrained_vae


# export GENERATED_TEXT_PATH=philly-dailydialog-epoch-2.0-beta-30.0
# export GENERATED_TEXT_FILE=../output/dialog/$GENERATED_TEXT_PATH/eval_text_generation_results.txt
# CUDA_VISIBLE_DEVICES=$GPU_ID python examples/big_ae/run_spacefusion_pretraining.py \
#     --dataset dailydialog \
#     --output_dir=../output/dialog/$GENERATED_TEXT_PATH \
#     --encoder_model_type=bert \
#     --encoder_model_name_or_path=bert-base-cased \
#     --decoder_model_type=gpt2 \
#     --decoder_model_name_or_path=gpt2 \
#     --train_data_file=$TRAIN_FILE \
#     --beta 2.0 \
#     --ratio_zero .5 \
#     --ratio_increase 0.25 \
#     --eval_data_file=$TEST_FILE \
#     --num_train_epochs 1.0 \
#     --save_steps 2000 \
#     --logging_steps 100 \
#     --overwrite_output_dir \
#     --per_gpu_train_batch_size 10 \
#     --block_size 512 \
#     --freeze_bert \
#     --per_gpu_eval_batch_size 1 \
#     --total_sents -1 \
#     --sents_per_cxt 10 \
#     --eval_generated_text_file $GENERATED_TEXT_FILE\
#     --checkpoint_dir ../output/dialog/$GENERATED_TEXT_PATH \
#     --gloabl_step_eval 10000  \
#     --use_pretrained_model \
#     --do_vis \
#     --path_ids=../data/datasets/dailydialog_data/dailydialog_data_1000.pt \
#     --n_pnt=64


#     --do_eval \

# export GENERATED_TEXT_FILE=../output/dialog/philly-dailydialog-epoch-5.0-beta-1.0/eval_text_generation_results.txt
# CUDA_VISIBLE_DEVICES=$GPU_ID python examples/big_ae/run_spacefusion_pretraining.py \
#     --dataset dailydialog \
#     --output_dir=../output/dialog/philly-dailydialog-epoch-5.0-beta-1.0 \
#     --encoder_model_type=bert \
#     --encoder_model_name_or_path=bert-base-cased \
#     --decoder_model_type=gpt2 \
#     --decoder_model_name_or_path=gpt2 \
#     --train_data_file=$TRAIN_FILE \
#     --do_eval \
#     --beta 2.0 \
#     --ratio_zero .5 \
#     --ratio_increase 0.25 \
#     --eval_data_file=$TEST_FILE \
#     --num_train_epochs 1.0 \
#     --save_steps 2000 \
#     --logging_steps 100 \
#     --overwrite_output_dir \
#     --per_gpu_train_batch_size 4 \
#     --block_size 512 \
#     --freeze_bert11 \
#     --per_gpu_eval_batch_size 1 \
#     --total_sents -1 \
#     --sents_per_cxt 10 \
#     --eval_generated_text_file $GENERATED_TEXT_FILE\
#     --checkpoint_dir ../output/dialog/philly-dailydialog-epoch-5.0-beta-1.0 \
#     --gloabl_step_eval 26000  \
#     --use_pretrained_model

# export GENERATED_TEXT_FILE=../output/dialog/philly-dailydialog-full-epoch-5.0-beta-1.0/eval_text_generation_results.txt
# CUDA_VISIBLE_DEVICES=$GPU_ID python examples/big_ae/run_spacefusion_pretraining.py \
#     --dataset dailydialog \
#     --output_dir=../output/dialog/philly-dailydialog-full-epoch-5.0-beta-1.0 \
#     --encoder_model_type=bert \
#     --encoder_model_name_or_path=bert-base-cased \
#     --decoder_model_type=gpt2 \
#     --decoder_model_name_or_path=gpt2 \
#     --train_data_file=$TRAIN_FILE \
#     --do_eval \
#     --beta 2.0 \
#     --ratio_zero .5 \
#     --ratio_increase 0.25 \
#     --eval_data_file=$TEST_FILE \
#     --num_train_epochs 1.0 \
#     --save_steps 2000 \
#     --logging_steps 100 \
#     --overwrite_output_dir \
#     --per_gpu_train_batch_size 4 \
#     --block_size 512 \
#     --freeze_bert11 \
#     --per_gpu_eval_batch_size 1 \
#     --total_sents -1 \
#     --sents_per_cxt 10 \
#     --eval_generated_text_file $GENERATED_TEXT_FILE \
#     --checkpoint_dir ../output/dialog/philly-dailydialog-full-epoch-5.0-beta-1.0 \
#     --gloabl_step_eval 26000  \
#     --use_pretrained_model

    #