Add ascend memory for pangu model
This commit is contained in:
parent
5491b2c982
commit
3211e527ad
|
@ -120,7 +120,7 @@ The vocab size of `vocab.vocab` is 40000, and the `eod id` is 6.
|
|||
|
||||
### Training On Ascend
|
||||
|
||||
Currently the scripts provide three default configures : `2.6B` `13B` and `200B`. The following command will start training `2.6B` model on 8 **Ascend cards**.
|
||||
Currently the scripts provide three default configures : `2.6B` `13B` and `200B`.
|
||||
|
||||
```bash
|
||||
|
||||
|
@ -465,7 +465,7 @@ Please check the official [homepage](https://gitee.com/mindspore/mindspore/tree/
|
|||
|
||||
# [Requirements](#contents)
|
||||
|
||||
- mindspore 1.2.1 or higher version
|
||||
- mindspore 1.3.0
|
||||
- jieba 0.42.1
|
||||
- sentencepiece 0.1.94
|
||||
- transformers >= 4.7.0
|
||||
|
|
|
@ -83,7 +83,7 @@ def run_train(args_opt):
|
|||
# Set execution mode
|
||||
context.set_context(mode=context.GRAPH_MODE,
|
||||
device_target=args_opt.device_target)
|
||||
context.set_context(variable_memory_max_size="30GB")
|
||||
context.set_context(variable_memory_max_size="31GB")
|
||||
# Set parallel context
|
||||
if args_opt.distribute == "true":
|
||||
D.init()
|
||||
|
|
Loading…
Reference in New Issue