Instructions to use pere/roberta-base-exp-32B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use pere/roberta-base-exp-32B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("fill-mask", model="pere/roberta-base-exp-32B")# Load model directly from transformers import AutoTokenizer, AutoModelForMaskedLM tokenizer = AutoTokenizer.from_pretrained("pere/roberta-base-exp-32B") model = AutoModelForMaskedLM.from_pretrained("pere/roberta-base-exp-32B") - Notebooks
- Google Colab
- Kaggle
Update run_mlm_flax_stream.py
Browse files- run_mlm_flax_stream.py +2 -2
run_mlm_flax_stream.py
CHANGED
|
@@ -449,8 +449,8 @@ if __name__ == "__main__":
|
|
| 449 |
|
| 450 |
# Store some constant
|
| 451 |
num_epochs = int(training_args.num_train_epochs)
|
| 452 |
-
train_batch_size = int(training_args.per_device_train_batch_size) * jax.device_count()
|
| 453 |
-
eval_batch_size = int(training_args.per_device_eval_batch_size) * jax.device_count()
|
| 454 |
|
| 455 |
# define number steps per stream epoch
|
| 456 |
num_train_steps = data_args.num_train_steps
|
|
|
|
| 449 |
|
| 450 |
# Store some constant
|
| 451 |
num_epochs = int(training_args.num_train_epochs)
|
| 452 |
+
train_batch_size = int(training_args.per_device_train_batch_size) * jax.device_count()
|
| 453 |
+
eval_batch_size = int(training_args.per_device_eval_batch_size) * jax.device_count()
|
| 454 |
|
| 455 |
# define number steps per stream epoch
|
| 456 |
num_train_steps = data_args.num_train_steps
|