File size: 43,574 Bytes
11389ae
f458b72
11389ae
 
 
 
 
 
 
 
 
 
 
 
 
 
f458b72
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c8b6302
f458b72
 
 
 
 
 
 
 
 
 
 
c8b6302
 
f458b72
 
 
c8b6302
f458b72
 
 
 
c8b6302
f458b72
 
 
 
c8b6302
f458b72
 
c8b6302
f458b72
 
c8b6302
f458b72
 
c8b6302
f458b72
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
========================
START TIME: Wed Jul  3 15:34:17 UTC 2024
python3 version = Python 3.10.14
========================
The token has not been saved to the git credentials helper. Pass `add_to_git_credential=True` in this function directly or `--add-to-git-credential` if using via `huggingface-cli` if you want to set the git credential as well.
Token is valid (permission: write).
Your token has been saved to /admin/home/ferdinand_mom/.cache/huggingface/token
Login successful
Already on 'bench_cluster'
M	examples/config_tiny_llama.py
M	examples/config_tiny_llama.yaml
M	examples/train_tiny_llama.sh
M	src/nanotron/models/llama.py
M	src/nanotron/trainer.py
Your branch is up to date with 'origin/bench_cluster'.
Job status: RUNNING
W0703 15:34:20.490000 139973111215936 torch/distributed/run.py:757] 
W0703 15:34:20.490000 139973111215936 torch/distributed/run.py:757] *****************************************
W0703 15:34:20.490000 139973111215936 torch/distributed/run.py:757] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
W0703 15:34:20.490000 139973111215936 torch/distributed/run.py:757] *****************************************
W0703 15:34:20.488000 140631068751680 torch/distributed/run.py:757] 
W0703 15:34:20.488000 140631068751680 torch/distributed/run.py:757] *****************************************
W0703 15:34:20.488000 140631068751680 torch/distributed/run.py:757] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
W0703 15:34:20.488000 140631068751680 torch/distributed/run.py:757] *****************************************
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config:
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config(general=GeneralArgs(project='bench_cluster',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            run='%date_%jobid',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            seed=42,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            step=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            consumed_train_samples=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            benchmark_csv_path=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            ignore_sanity_checks=True),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        parallelism=ParallelismArgs(dp=2,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    pp=8,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    tp=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    pp_engine=<nanotron.parallel.pipeline_parallel.engine.OneForwardOneBackwardPipelineEngine object at 0x7fd00c7dc910>,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    tp_mode=<TensorParallelLinearMode.REDUCE_SCATTER: 2>,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    tp_linear_async_communication=False,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    expert_parallel_size=1),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        model=ModelArgs(model_config=LlamaConfig(bos_token_id=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 eos_token_id=2,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 hidden_act='silu',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 hidden_size=2048,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 initializer_range=0.02,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 intermediate_size=4096,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 is_llama_config=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 max_position_embeddings=4096,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 num_attention_heads=32,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 num_hidden_layers=24,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 num_key_value_heads=32,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 pad_token_id=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 pretraining_tp=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 rms_norm_eps=1e-05,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 rope_scaling=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 rope_theta=10000.0,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 tie_word_embeddings=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 use_cache=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                 vocab_size=50257),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                        init_method=RandomInit(std=0.025),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                        dtype=torch.bfloat16,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                        make_vocab_size_divisible_by=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                        ddp_bucket_cap_mb=25),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        tokenizer=TokenizerArgs(tokenizer_name_or_path='openai-community/gpt2',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                tokenizer_revision=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                tokenizer_max_length=None),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        checkpoints=CheckpointsArgs(checkpoints_path=Path('/dev/null'),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    checkpoint_interval=100000,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    save_initial_state=False,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    resume_checkpoint_path=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                    checkpoints_path_is_shared_file_system=False),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        logging=LoggingArgs(log_level='info',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            log_level_replica='info',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                            iteration_step_info_interval=1),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        tokens=TokensArgs(sequence_length=4096,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                          train_steps=20,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                          micro_batch_size=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                          batch_accumulation_per_replica=512,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                          val_check_interval=-1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                          limit_val_batches=0,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                          limit_test_batches=0),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                     adam_beta1=0.9,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                     adam_beta2=0.95,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                     torch_adam_is_fused=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                     name='adamW'),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                zero_stage=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                weight_decay=0.01,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                clip_grad=1.0,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                accumulate_grad_in_fp32=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0001,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                        lr_warmup_steps=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                        lr_warmup_style='linear',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                        lr_decay_style='linear',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                        lr_decay_steps=19,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                        lr_decay_starting_step=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                        min_decay_lr=1e-05)),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        data_stages=[DatasetStageArgs(name='Training Stage',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                      start_training_step=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                      data=DataArgs(dataset=PretrainDatasetsArgs(hf_dataset_or_datasets='roneneldan/TinyStories',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                                 hf_dataset_splits='train',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                                 hf_dataset_config_name=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                                 dataset_processing_num_proc_per_process=64,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                                 dataset_overwrite_cache=False,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                                                 text_column_name='text'),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                    seed=42,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:                                                    num_loading_workers=32))],
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        profiler=ProfilerArgs(profiler_export_path=Path('/fsx/ferdinandmom/ferdinand-hf/bench_cluster/results/llama-1B/16_GPUS/dp-2_tp-1_pp-8_mbz-1')),
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:        lighteval=None)
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Model Config:
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: LlamaConfig(bos_token_id=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             eos_token_id=2,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             hidden_act='silu',
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             hidden_size=2048,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             initializer_range=0.02,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             intermediate_size=4096,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             is_llama_config=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             max_position_embeddings=4096,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             num_attention_heads=32,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             num_hidden_layers=24,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             num_key_value_heads=32,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             pad_token_id=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             pretraining_tp=1,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             rms_norm_eps=1e-05,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             rope_scaling=None,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             rope_theta=10000.0,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             tie_word_embeddings=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             use_cache=True,
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:             vocab_size=50257)
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Building model..
[default0]:07/03/2024 15:34:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Setting PP block ranks...
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Total number of parameters: 1.21G (2312.82MiB)
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Local number of parameters: 271M (516.35MiB)
[default4]:07/03/2024 15:34:51 [INFO|DP=0|PP=2|TP=0|ip-26-0-160-225]: Local number of parameters: 126M (240.02MiB)
[default4]:07/03/2024 15:34:51 [INFO|DP=0|PP=2|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 243.03MiB. Peak allocated: 245.06MiB Peak reserved: 262.00MiB
[default4]:07/03/2024 15:34:51 [INFO|DP=0|PP=2|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default6]:07/03/2024 15:34:51 [INFO|DP=0|PP=3|TP=0|ip-26-0-160-225]: Local number of parameters: 168M (320.03MiB)
[default6]:07/03/2024 15:34:51 [INFO|DP=0|PP=3|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 324.04MiB. Peak allocated: 326.07MiB Peak reserved: 336.00MiB
[default6]:07/03/2024 15:34:51 [INFO|DP=0|PP=3|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default2]:07/03/2024 15:34:51 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: Local number of parameters: 126M (240.02MiB)
[default2]:07/03/2024 15:34:51 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 243.03MiB. Peak allocated: 245.06MiB Peak reserved: 262.00MiB
[default2]:07/03/2024 15:34:51 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default6]:07/03/2024 15:34:51 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: Local number of parameters: 103M (196.32MiB)
[default6]:07/03/2024 15:34:51 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 196.33MiB. Peak allocated: 196.34MiB Peak reserved: 200.00MiB
[default6]:07/03/2024 15:34:51 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default4]:07/03/2024 15:34:51 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: Local number of parameters: 168M (320.03MiB)
[default4]:07/03/2024 15:34:51 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 324.04MiB. Peak allocated: 326.07MiB Peak reserved: 336.00MiB
[default4]:07/03/2024 15:34:51 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: Local number of parameters: 126M (240.02MiB)
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 243.03MiB. Peak allocated: 245.06MiB Peak reserved: 262.00MiB
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default2]:07/03/2024 15:34:51 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: Local number of parameters: 126M (240.02MiB)
[default2]:07/03/2024 15:34:51 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 243.03MiB. Peak allocated: 245.06MiB Peak reserved: 262.00MiB
[default2]:07/03/2024 15:34:51 [INFO|DP=0|PP=1|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 520.36MiB. Peak allocated: 522.39MiB Peak reserved: 534.00MiB
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default0]:07/03/2024 15:34:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Parametrizing model parameters using StandardParametrizator
[default7]:07/03/2024 15:34:52 [INFO|DP=1|PP=7|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default1]:07/03/2024 15:34:52 [INFO|DP=1|PP=4|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default3]:07/03/2024 15:34:52 [INFO|DP=1|PP=5|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default5]:07/03/2024 15:34:52 [INFO|DP=1|PP=6|TP=0|ip-26-0-170-160]: No checkpoint path provided.
[default1]:07/03/2024 15:34:52 [INFO|DP=1|PP=0|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default3]:07/03/2024 15:34:52 [INFO|DP=1|PP=1|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default5]:07/03/2024 15:34:52 [INFO|DP=1|PP=2|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default7]:07/03/2024 15:34:52 [INFO|DP=1|PP=3|TP=0|ip-26-0-160-225]: No checkpoint path provided.
[default0]:07/03/2024 15:34:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Optimizer Building] Using LearningRateForSP as learning rate
[default0]:07/03/2024 15:34:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] Size of optimizer params per rank:
[default0]:07/03/2024 15:34:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 0 has 135M out of 271M (50.00%) params' optimizer states
[default0]:07/03/2024 15:34:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 1 has 135M out of 271M (50.00%) params' optimizer states
[default0]:07/03/2024 15:34:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] Stage Training Stage has 19 remaining training steps and has consumed 0 samples
[default0]:07/03/2024 15:34:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Using `datasets` library
[default0]:07/03/2024 15:34:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Loading tokenizer from openai-community/gpt2 and transformers/hf_hub versions ('4.41.2', '0.23.4')
[default0]:07/03/2024 15:34:56 [WARNING|DP=0|PP=0|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default0]:Repo card metadata block was not found. Setting CardData to empty.
[default0]:07/03/2024 15:34:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] There are 1 training stages 
[default0]:07/03/2024 15:34:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Stage Training Stage] start from step 1 
[default0]:07/03/2024 15:34:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: 
[default0]:07/03/2024 15:34:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Start training] datetime: 2024-07-03 15:34:57.136501 | mbs: 1 | grad_accum: 512 | global_batch_size: 1024 | sequence_length: 4096 | train_steps: 20 | start_iteration_step: 0 | consumed_train_samples: 0
[default0]:07/03/2024 15:34:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Resuming training from stage Training Stage, it has trained for 0 samples and has 19 remaining train steps
[default0]:07/03/2024 15:34:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:  Memory usage: 2069.40MiB. Peak allocated 2069.40MiB. Peak reserved: 2086.00MiB
[default6]:07/03/2024 15:34:57 [WARNING|DP=0|PP=7|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default7]:07/03/2024 15:34:57 [WARNING|DP=1|PP=7|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default4]:07/03/2024 15:34:57 [WARNING|DP=0|PP=6|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default7]:Repo card metadata block was not found. Setting CardData to empty.
[default5]:07/03/2024 15:34:57 [WARNING|DP=1|PP=6|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default5]:Repo card metadata block was not found. Setting CardData to empty.
[default4]:Repo card metadata block was not found. Setting CardData to empty.
[default6]:Repo card metadata block was not found. Setting CardData to empty.
[default4]:07/03/2024 15:34:57 [WARNING|DP=0|PP=2|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default5]:07/03/2024 15:34:57 [WARNING|DP=1|PP=2|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default7]:Repo card metadata block was not found. Setting CardData to empty.
[default7]:07/03/2024 15:34:57 [WARNING|DP=1|PP=3|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default6]:07/03/2024 15:34:57 [WARNING|DP=0|PP=3|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default5]:Repo card metadata block was not found. Setting CardData to empty.
[default4]:Repo card metadata block was not found. Setting CardData to empty.
[default6]:Repo card metadata block was not found. Setting CardData to empty.
[default2]:07/03/2024 15:34:57 [WARNING|DP=0|PP=5|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default1]:07/03/2024 15:34:57 [WARNING|DP=1|PP=4|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default3]:07/03/2024 15:34:57 [WARNING|DP=1|PP=5|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default0]:07/03/2024 15:34:57 [WARNING|DP=0|PP=4|TP=0|ip-26-0-170-160]: Repo card metadata block was not found. Setting CardData to empty.
[default2]:Repo card metadata block was not found. Setting CardData to empty.
[default0]:Repo card metadata block was not found. Setting CardData to empty.
[default3]:Repo card metadata block was not found. Setting CardData to empty.
[default1]:Repo card metadata block was not found. Setting CardData to empty.
[default1]:07/03/2024 15:34:57 [WARNING|DP=1|PP=0|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default3]:07/03/2024 15:34:57 [WARNING|DP=1|PP=1|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default3]:Repo card metadata block was not found. Setting CardData to empty.
[default1]:Repo card metadata block was not found. Setting CardData to empty.
[default2]:07/03/2024 15:34:57 [WARNING|DP=0|PP=1|TP=0|ip-26-0-160-225]: Repo card metadata block was not found. Setting CardData to empty.
[default2]:Repo card metadata block was not found. Setting CardData to empty.
[default7]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default7]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default5]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default5]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default3]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default3]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default1]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default1]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default7]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default7]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default5]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default5]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default3]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default3]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default1]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default1]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default6]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default6]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default7]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions
[default7]:  warnings.warn(
[default4]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default4]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default1]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions
[default1]:  warnings.warn(
[default2]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default2]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default0]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: Attempting to run cuBLAS, but there was no current CUDA context! Attempting to set the primary context... (Triggered internally at ../aten/src/ATen/cuda/CublasHandlePool.cpp:135.)
[default0]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default0]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default0]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default6]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default6]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default4]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default4]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default2]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default2]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default0]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: Attempting to run cuBLAS, but there was no current CUDA context! Attempting to set the primary context... (Triggered internally at ../aten/src/ATen/cuda/CublasHandlePool.cpp:135.)
[default0]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default0]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: c10d::allreduce_: an autograd kernel was not registered to the Autograd key(s) but we are trying to backprop through it. This may lead to silently incorrect behavior. This behavior is deprecated and will be removed in a future version of PyTorch. If your operator is differentiable, please ensure you have registered an autograd kernel to the correct Autograd key (e.g. DispatchKey::Autograd, DispatchKey::CompositeImplicitAutograd). If your operator is not differentiable, or to squash this warning and use the previous behavior, please register torch::CppFunction::makeFallthrough() to DispatchKey::Autograd. (Triggered internally at ../torch/csrc/autograd/autograd_not_implemented_fallback.cpp:63.)
[default0]:  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[default6]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions
[default6]:  warnings.warn(
[default0]:07/03/2024 15:36:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:  Memory usage: 2135.44MiB. Peak allocated 11566.15MiB. Peak reserved: 11748.00MiB
[default0]:/fsx/ferdinandmom/miniforge3/envs/env-bench-cluster/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py:2261: UserWarning: torch.distributed.all_reduce_coalesced will be deprecated. If you must use it, please revisit our documentation later at https://pytorch.org/docs/master/distributed.html#collective-functions
[default0]:  warnings.warn(
[default6]:07/03/2024 15:36:21 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: iteration: 1 / 20 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 83.6K | tokens_per_sec: 50.2K | tokens_per_sec_per_gpu: 3.14K | global_batch_size: 1.02K | lm_loss: 11.1 | lr: 0.0001 | model_tflops_per_gpu: 28.5 | hardware_tflops_per_gpu: 28.5 | grad_norm: 24.9 | cuda_memory_allocated: 1.3G | cuda_max_memory_reserved: 4.01G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G
[default0]:07/03/2024 15:36:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]:  Memory usage: 3168.14MiB. Peak allocated 4459.01MiB. Peak reserved: 13244.00MiB