|
--- |
|
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). |
|
|
|
## ALBERT XXLarge model HPU configuration |
|
|
|
This model only contains the `GaudiConfig` file for running the [albert-xxlarge-v1](https://huggingface.co/albert-xxlarge-v1) 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_Mixed_Precision/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. |
|
|
|
[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 ALBERT XXL with the following command: |
|
```bash |
|
python run_qa.py \ |
|
--model_name_or_path albert-xxlarge-v1 \ |
|
--gaudi_config_name Habana/albert-xxlarge-v1 \ |
|
--dataset_name squad \ |
|
--do_train \ |
|
--do_eval \ |
|
--per_device_train_batch_size 12 \ |
|
--per_device_eval_batch_size 2 \ |
|
--learning_rate 5e-6 \ |
|
--num_train_epochs 2 \ |
|
--max_seq_length 384 \ |
|
--output_dir /tmp/squad/ \ |
|
--use_habana \ |
|
--use_lazy_mode \ |
|
--throughput_warmup_steps 2 |
|
``` |
|
|
|
Check the [documentation](https://huggingface.co/docs/optimum/habana/index) out for more advanced usage and examples. |
