Spaces:
Running
Running
Merge pull request #24 from borisdayma/feat--log-model-frequently
Browse files- seq2seq/run_seq2seq_flax.py +44 -29
seq2seq/run_seq2seq_flax.py
CHANGED
@@ -84,7 +84,7 @@ MODEL_TYPES = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES)
|
|
84 |
OUTPUT_VOCAB_SIZE = 16384 + 1 # encoded image token space + 1 for bos
|
85 |
OUTPUT_LENGTH = 256 + 1 # number of encoded tokens + 1 for bos
|
86 |
BOS_TOKEN_ID = 16384
|
87 |
-
BASE_MODEL = 'facebook/bart-large'
|
88 |
|
89 |
|
90 |
@dataclass
|
@@ -231,6 +231,12 @@ class DataTrainingArguments:
|
|
231 |
log_model: bool = field(
|
232 |
default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
|
233 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
234 |
|
235 |
def __post_init__(self):
|
236 |
if self.dataset_name is None and self.train_file is None and self.validation_file is None:
|
@@ -340,7 +346,7 @@ def wandb_log(metrics, step=None, prefix=None):
|
|
340 |
if jax.process_index() == 0:
|
341 |
log_metrics = {f'{prefix}/{k}' if prefix is not None else k: jax.device_get(v) for k,v in metrics.items()}
|
342 |
if step is not None:
|
343 |
-
log_metrics
|
344 |
wandb.log(log_metrics)
|
345 |
|
346 |
|
@@ -773,6 +779,38 @@ def main():
|
|
773 |
|
774 |
return eval_metrics
|
775 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
776 |
for epoch in epochs:
|
777 |
# ======================== Training ================================
|
778 |
train_start = time.time()
|
@@ -795,6 +833,9 @@ def main():
|
|
795 |
|
796 |
if global_step % training_args.eval_steps == 0:
|
797 |
run_evaluation()
|
|
|
|
|
|
|
798 |
|
799 |
# log final train metrics
|
800 |
wandb_log(unreplicate(train_metric), step=global_step, prefix='train')
|
@@ -809,34 +850,8 @@ def main():
|
|
809 |
eval_metrics = run_evaluation()
|
810 |
|
811 |
# save checkpoint after each epoch and push checkpoint to the hub
|
812 |
-
|
813 |
-
params = jax.device_get(jax.tree_map(lambda x: x[0], state.params))
|
814 |
-
|
815 |
-
# save model locally
|
816 |
-
model.save_pretrained(
|
817 |
-
training_args.output_dir,
|
818 |
-
params=params,
|
819 |
-
)
|
820 |
-
|
821 |
-
# save to W&B
|
822 |
-
if data_args.log_model:
|
823 |
-
metadata = {'epoch': epoch+1, 'eval/loss': eval_metrics['loss']}
|
824 |
-
artifact = wandb.Artifact(
|
825 |
-
name=f"model-{wandb.run.id}", type="bart_model", metadata=metadata
|
826 |
-
)
|
827 |
-
artifact.add_file(str(Path(training_args.output_dir) / 'flax_model.msgpack'))
|
828 |
-
artifact.add_file(str(Path(training_args.output_dir) / 'config.json'))
|
829 |
-
wandb.run.log_artifact(artifact)
|
830 |
|
831 |
-
# save to the hub
|
832 |
-
if training_args.push_to_hub:
|
833 |
-
model.save_pretrained(
|
834 |
-
training_args.output_dir,
|
835 |
-
params=params,
|
836 |
-
push_to_hub=training_args.push_to_hub,
|
837 |
-
commit_message=f"Saving weights and logs of epoch {epoch+1}",
|
838 |
-
temp_dir=True # avoid issues with being in a repository
|
839 |
-
)
|
840 |
|
841 |
# ======================== Prediction loop ==============================
|
842 |
if training_args.do_predict:
|
|
|
84 |
OUTPUT_VOCAB_SIZE = 16384 + 1 # encoded image token space + 1 for bos
|
85 |
OUTPUT_LENGTH = 256 + 1 # number of encoded tokens + 1 for bos
|
86 |
BOS_TOKEN_ID = 16384
|
87 |
+
BASE_MODEL = 'facebook/bart-large-cnn' # we currently have issues with bart-large
|
88 |
|
89 |
|
90 |
@dataclass
|
|
|
231 |
log_model: bool = field(
|
232 |
default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
|
233 |
)
|
234 |
+
save_model_steps: Optional[int] = field(
|
235 |
+
default=3000, # about once every hour in our experiments
|
236 |
+
metadata={
|
237 |
+
"help": "For logging the model more frequently. Used only when `log_model` is set."
|
238 |
+
},
|
239 |
+
)
|
240 |
|
241 |
def __post_init__(self):
|
242 |
if self.dataset_name is None and self.train_file is None and self.validation_file is None:
|
|
|
346 |
if jax.process_index() == 0:
|
347 |
log_metrics = {f'{prefix}/{k}' if prefix is not None else k: jax.device_get(v) for k,v in metrics.items()}
|
348 |
if step is not None:
|
349 |
+
log_metrics['train/step'] = step
|
350 |
wandb.log(log_metrics)
|
351 |
|
352 |
|
|
|
779 |
|
780 |
return eval_metrics
|
781 |
|
782 |
+
def run_save_model(step, epoch, eval_metrics=None):
|
783 |
+
if jax.process_index() == 0:
|
784 |
+
params = jax.device_get(jax.tree_map(lambda x: x[0], state.params))
|
785 |
+
|
786 |
+
# save model locally
|
787 |
+
model.save_pretrained(
|
788 |
+
training_args.output_dir,
|
789 |
+
params=params,
|
790 |
+
)
|
791 |
+
|
792 |
+
# save to W&B
|
793 |
+
if data_args.log_model:
|
794 |
+
metadata = {'step': step, 'epoch': epoch}
|
795 |
+
if eval_metrics is not None:
|
796 |
+
metadata['eval/loss'] = eval_metrics['loss']
|
797 |
+
artifact = wandb.Artifact(
|
798 |
+
name=f"model-{wandb.run.id}", type="bart_model", metadata=metadata
|
799 |
+
)
|
800 |
+
artifact.add_file(str(Path(training_args.output_dir) / 'flax_model.msgpack'))
|
801 |
+
artifact.add_file(str(Path(training_args.output_dir) / 'config.json'))
|
802 |
+
wandb.run.log_artifact(artifact)
|
803 |
+
|
804 |
+
# save to the hub
|
805 |
+
if training_args.push_to_hub:
|
806 |
+
model.save_pretrained(
|
807 |
+
training_args.output_dir,
|
808 |
+
params=params,
|
809 |
+
push_to_hub=training_args.push_to_hub,
|
810 |
+
commit_message=f"Saving weights and logs of epoch {epoch+1}",
|
811 |
+
temp_dir=True # avoid issues with being in a repository
|
812 |
+
)
|
813 |
+
|
814 |
for epoch in epochs:
|
815 |
# ======================== Training ================================
|
816 |
train_start = time.time()
|
|
|
833 |
|
834 |
if global_step % training_args.eval_steps == 0:
|
835 |
run_evaluation()
|
836 |
+
|
837 |
+
if global_step % data_args.save_model_steps == 0:
|
838 |
+
run_save_model(global_step, epoch)
|
839 |
|
840 |
# log final train metrics
|
841 |
wandb_log(unreplicate(train_metric), step=global_step, prefix='train')
|
|
|
850 |
eval_metrics = run_evaluation()
|
851 |
|
852 |
# save checkpoint after each epoch and push checkpoint to the hub
|
853 |
+
run_save_model(global_step, epoch, eval_metrics)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
854 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
855 |
|
856 |
# ======================== Prediction loop ==============================
|
857 |
if training_args.do_predict:
|