# Explicitly set number of training steps using Trainer
**URL:**
**Author:**  [@aclifton314](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Faclifton314)%5C
**Post date:** [September 14, 2020, 7:11pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fexplicitly-set-number-of-training-steps-using-trainer%2F1127%2F1 "2020-09-14T19:11:17Z")
I’m using `Trainer` to handle finetuning a GPT2 model. I see in `TrainingArguments` there is a `max_steps` that overrides `num_train_epochs`.
For a batch size of 32, is setting `max_steps=1000000` the equivalent of setting `num_train_epochs=31250`?
Also, what happens if I have a batch size of 6 but want to set `max_steps=1000000`? Does `Trainer` stop training after the nearest divisible whole number or does it change the batch size at the end?
Thanks in advance!!
---
**Author:**  [@sgugger](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsgugger)%5C
**Post date:** [September 14, 2020, 9:00pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fexplicitly-set-number-of-training-steps-using-trainer%2F1127%2F2 "2020-09-14T21:00:14Z")
The number of steps is the number of update steps. It’s not the number of training examples to be seen.
---
**Author:**  [@aclifton314](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Faclifton314)%5C
**Post date:** [September 14, 2020, 9:03pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fexplicitly-set-number-of-training-steps-using-trainer%2F1127%2F3 "2020-09-14T21:03:51Z")
Ok. Is it then the case that for a batch size of 32, setting `max_steps=1000000` is the same as setting `num_train_epochs=31250` ?
---
**Author:**  [@sgugger](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsgugger)%5C
**Post date:** [September 15, 2020, 12:01pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fexplicitly-set-number-of-training-steps-using-trainer%2F1127%2F4 "2020-09-15T12:01:41Z")
No (unless your dataset has 32 batches so is of length 32\*32). `num_train_epochs = max_steps / len(train_dataloader)` in general.
---
**Author:**  [@aclifton314](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Faclifton314)%5C
**Post date:** [September 15, 2020, 4:45pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fexplicitly-set-number-of-training-steps-using-trainer%2F1127%2F5 "2020-09-15T16:45:05Z")
Apologies for my confusion. Is then `max_steps` the maximum number of times a batch makes a forward and backward pass through the network?
---
**Author:**  [@sgugger](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsgugger)%5C
**Post date:** [September 16, 2020, 1:06am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fexplicitly-set-number-of-training-steps-using-trainer%2F1127%2F6 "2020-09-16T01:06:13Z")
Not exactly it is the number of update steps (so if you’re using gradient accumulation, it’s a bit different than just doing forward and backward).