File size: 2,109 Bytes
578c3d0
 
 
9a302a2
 
e00aa88
9a302a2
84bbd47
9a302a2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
63d8b31
9a302a2
 
63d8b31
9a302a2
 
 
 
 
 
 
 
63d8b31
9a302a2
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
---

[Optimum Habana](https://github.com/huggingface/optimum-habana) is the interface between the Transformers library and Habana's Gaudi processor (HPU). It provides a set of tools enabling easy and fast model loading and fine-tuning on single- and multi-HPU settings for different downstream tasks.
Learn more about how to take advantage of the power of Habana HPUs to train Transformers models at [hf.co/hardware/habana](https://huggingface.co/hardware/habana).

## DistilBERT Base model HPU configuration

This model contains just the `GaudiConfig` file for running the [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) model on Habana's Gaudi processors (HPU).

**This model contains no model weights, only a GaudiConfig.**

This enables to specify:
- `use_habana_mixed_precision`: whether to use Habana Mixed Precision (HMP)
    - `hmp_opt_level`: optimization level for HMP, see [here](https://docs.habana.ai/en/latest/PyTorch/PyTorch_User_Guide/PT_Mixed_Precision.html#configuration-options) for a detailed explanation
    - `hmp_bf16_ops`: list of operators that should run in bf16
    - `hmp_fp32_ops`: list of operators that should run in fp32
    - `hmp_is_verbose`: verbosity
- `use_fused_adam`: whether to use Habana's custom AdamW implementation
- `use_fused_clip_norm`: whether to use Habana's fused gradient norm clipping operator

## Usage

The model is instantiated the same way as in the Transformers library.
The only difference is that there are a few new training arguments specific to HPUs:

```
from optimum.habana import GaudiTrainer, GaudiTrainingArguments
from transformers import DistilBertTokenizer, DistilBertModel

tokenizer = DistilBertTokenizer.from_pretrained("distilbert-base-uncased")
model = DistilBertModel.from_pretrained("distilbert-base-uncased")
args = GaudiTrainingArguments(
    output_dir="/tmp/output_dir",
    use_habana=True,
    use_lazy_mode=True,
    gaudi_config_name="Habana/distilbert-base-uncased",
)

trainer = GaudiTrainer(
    model=model,
    args=args,
    tokenizer=tokenizer,
)
trainer.train()
```