protnote/amlt_config.yml at main · microsoft/protnote · GitHub

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
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
# For field details: https://amulet-docs.azurewebsites.net/config_file.html
description: Test ddp #Normal training with augmentations part2 v3

target:
  service: sing
  name: barlow01
  workspace_name: bio0-ext
environment:
  image: amlt-sing/acpt-2.0.1-py3.10-cuda11.8 #azureml/openmpi4.1.0-cuda11.8-cudnn8-ubuntu22.04:latest
  conda_yaml_file: $CONFIG_DIR/environment.yml
  skip_conda_packages_on_sing:
  - tensorflow
  - cudatoolkit
  - deepspeed
  - pip
  - python\b


code:
  local_dir: $CONFIG_DIR/

data:
  local_dir: $CONFIG_DIR/data
  remote_dir: data

# SKU usage: G1 (single GPU), G4 (quad GPU), G4-V100 (1 machine, 4 V100 gpus), etc...
jobs:
# - name: '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}{weighted_sampling False}'
#   sku: G8
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH_ZERO_SHOT
#     --validation-path-name VAL_DATA_PATH_ZERO_SHOT
#     --test-paths-names UNSEEN_DATA_PATH_ZERO_SHOT
#     --name '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}{weighted_sampling False}'
#     --override EXTRACT_VOCABULARIES_FROM null ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label WEIGHTED_SAMPLING False
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"
# - name: '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}{loss_fn BCE}'
#   sku: G8
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH_ZERO_SHOT
#     --validation-path-name VAL_DATA_PATH_ZERO_SHOT
#     --test-paths-names UNSEEN_DATA_PATH_ZERO_SHOT
#     --name '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}{loss_fn BCE}'
#     --override EXTRACT_VOCABULARIES_FROM null ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label LOSS_FN BCE
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"
# - name: '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder biogpt}'
#   sku: G8
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH_ZERO_SHOT
#     --validation-path-name VAL_DATA_PATH_ZERO_SHOT
#     --test-paths-names UNSEEN_DATA_PATH_ZERO_SHOT
#     --name '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder biogpt}'
#     --override EXTRACT_VOCABULARIES_FROM null ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT microsoft/biogpt AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"
# - name: '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 0}{label_encoder E5_multiling_inst}'
#   sku: G8
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH_ZERO_SHOT
#     --validation-path-name VAL_DATA_PATH_ZERO_SHOT
#     --test-paths-names UNSEEN_DATA_PATH_ZERO_SHOT
#     --name '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 0}{label_encoder E5_multiling_inst}'
#     --override EXTRACT_VOCABULARIES_FROM null ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 0 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"
# - name: '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0}{label_noise 20}{label_encoder E5_multiling_inst}'
#   sku: G8
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH_ZERO_SHOT
#     --validation-path-name VAL_DATA_PATH_ZERO_SHOT
#     --test-paths-names UNSEEN_DATA_PATH_ZERO_SHOT
#     --name '{zero_shot}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0}{label_noise 20}{label_encoder E5_multiling_inst}'
#     --override EXTRACT_VOCABULARIES_FROM null ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0 LABEL_EMBEDDING_NOISING_ALPHA 20 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"


- name: 'test_ddp'
  sku: G8-V100
  priority: high
  preemptible: False
  command:
  - python main.py
    --nodes 1
    --gpus 8
    --train-path-name TRAIN_DATA_PATH
    --validation-path-name VAL_DATA_PATH
    --name "test_ddp"
    --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 3 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label TRAIN_SUBSET_FRACTION 0.01 VALIDATION_SUBSET_FRACTION 0.1
    --amlt
    --use-wandb
  submit_args:
    env:
      WANDB_BASE_URL: "https://microsoft-research.wandb.io"
      WANDB_API_KEY: "$WANDB_API_KEY"


# - name: '{normal_clustered}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#   sku: G8-V100
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_CLUSTERED_DATA_PATH
#     --validation-path-name VAL_CLUSTERED_DATA_PATH
#     --name '{normal_clustered}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#     --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 100 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"
# - name: '{normal}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#   sku: G8-V100
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH
#     --validation-path-name VAL_DATA_PATH
#     --name '{normal}{MLP Large}{descriptions names+label}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#     --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 100 LABEL_AUGMENTATION_DESCRIPTIONS name+label INFERENCE_GO_DESCRIPTIONS name+label
#     --amlt
#     --use-wandb
#     --load-model 'models/ProTCL/normal_test_label_aug_v4.pt'
#     --from-checkpoint
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"
# - name: '{normal}{MLP Large}{descriptions names+label+synonyms}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#   sku: G8-V100
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH
#     --validation-path-name VAL_DATA_PATH
#     --name '{normal}{MLP Large}{descriptions names+label+synonyms}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#     --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 100 LABEL_AUGMENTATION_DESCRIPTIONS name+label+synonym_exact INFERENCE_GO_DESCRIPTIONS name+label
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"
# - name: '{zero_shot}{MLP Large}{descriptions names+label+synonyms}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#   sku: G8
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH_ZERO_SHOT
#     --validation-path-name VAL_DATA_PATH_ZERO_SHOT
#     --name '{zero_shot}{MLP Large}{descriptions names+label+synonyms}{ensemble True}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#     --override EXTRACT_VOCABULARIES_FROM null ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 LABEL_AUGMENTATION_DESCRIPTIONS name+label+synonym_exact INFERENCE_GO_DESCRIPTIONS name+label
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"


# - name: '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#   sku: G8-V100
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH
#     --validation-path-name VAL_DATA_PATH
#     --test-paths-names TEST_DATA_PATH
#     --name '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 20}{label_encoder E5_multiling_inst}'
#     --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 20 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 64
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"

# - name: '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 10}{label_encoder E5_multiling_inst}'
#   sku: G8-V100
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH
#     --validation-path-name VAL_DATA_PATH
#     --test-paths-names TEST_DATA_PATH
#     --name '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 10}{label_encoder E5_multiling_inst}'
#     --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 10 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 64
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"

# - name: '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 0}{label_encoder E5_multiling_inst}'
#   sku: G8-V100
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH
#     --validation-path-name VAL_DATA_PATH
#     --test-paths-names TEST_DATA_PATH
#     --name '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 0}{label_encoder E5_multiling_inst}'
#     --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT intfloat/multilingual-e5-large-instruct AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 0 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 64
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"

# - name: '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 0}{label_encoder biogpt}'
#   sku: G8-V100
#   priority: high
#   preemptible: False
#   command:
#   - python main.py
#     --nodes 1
#     --gpus 8
#     --train-path-name TRAIN_DATA_PATH
#     --validation-path-name VAL_DATA_PATH
#     --test-paths-names TEST_DATA_PATH
#     --name '{normal}{MLP Large}{descriptions names}{pooling mean}{sequence_augmentations 0.1}{label_noise 0}{label_encoder biogpt}'
#     --override ESTIMATE_MAP True OPTIMIZATION_METRIC_NAME f1_macro LABEL_ENCODER_CHECKPOINT microsoft/biogpt AUGMENT_RESIDUE_PROBABILITY 0.1 LABEL_EMBEDDING_NOISING_ALPHA 0 VALIDATION_BATCH_SIZE 8 TEST_BATCH_SIZE 8 NUM_EPOCHS 64
#     --amlt
#     --use-wandb
#   submit_args:
#     env:
#       WANDB_BASE_URL: "https://microsoft-research.wandb.io"
#       WANDB_API_KEY: "$WANDB_API_KEY"