File size: 2,021 Bytes
2c9a1ec
 
 
7ce8c0c
fd433d3
 
 
7ce8c0c
 
 
6b2354a
7ce8c0c
 
 
 
98feec4
7ce8c0c
 
 
 
 
 
0c115bd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
de83b49
 
7ce8c0c
0c115bd
 
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
---
license: apache-2.0
---

[Optimum Habana](https://github.com/huggingface/optimum-habana) is the interface between the Hugging Face Transformers and Diffusers libraries and Habana's Gaudi processor (HPU).
It provides a set of tools enabling easy and fast model loading, training and inference 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 and deploy Transformers and Diffusers models at [hf.co/hardware/habana](https://huggingface.co/hardware/habana).

## RoBERTa Large model HPU configuration

This model only contains the `GaudiConfig` file for running the [roberta-large](https://huggingface.co/roberta-large) model on Habana's Gaudi processors (HPU).

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

This enables to specify:
- `use_torch_autocast`: whether to use PyTorch's autocast mixed precision
- `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.

[Here](https://github.com/huggingface/optimum-habana/blob/main/examples/question-answering/run_qa.py) is a question-answering example script to fine-tune a model on SQuAD. You can run it with RoBERTa Large with the following command:
```bash
python run_qa.py \
  --model_name_or_path roberta-large \
  --gaudi_config_name Habana/roberta-large \
  --dataset_name squad \
  --do_train \
  --do_eval \
  --per_device_train_batch_size 12 \
  --per_device_eval_batch_size 8 \
  --learning_rate 3e-5 \
  --num_train_epochs 2 \
  --max_seq_length 384 \
  --output_dir /tmp/squad/ \
  --use_habana \
  --use_lazy_mode \
  --throughput_warmup_steps 3 \
  --bf16
```

Check the [documentation](https://huggingface.co/docs/optimum/habana/index) out for more advanced usage and examples.