<!---
Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->
![](https://github.com/huggingface/optimum-habana/blob/main/readme_logo.png)
# Optimum Habana
🤗 Optimum Habana is the interface between the 🤗 Transformers and Diffusers libraries and [Habana's Gaudi processor (HPU)](https://docs.habana.ai/en/latest/index.html).
It provides a set of tools enabling easy model loading, training and inference on single- and multi-HPU settings for different downstream tasks.
The list of officially validated models and tasks is available [here](https://github.com/huggingface/optimum-habana#validated-models). Users can try other models and tasks with only few changes.
## What is a Habana Processing Unit (HPU)?
HPUs offer fast model training and inference as well as a great price-performance ratio.
Check out [this blog post about BERT pre-training](https://huggingface.co/blog/pretraining-bert) and [this article benchmarking Habana Gaudi2 versus Nvidia A100 GPUs](https://huggingface.co/blog/habana-gaudi-2-benchmark) for concrete examples.
If you are not familiar with HPUs and would like to know more about them, we recommend you take a look at [our conceptual guide](https://huggingface.co/docs/optimum/habana/concept_guides/hpu).
## Install the library and get example scripts
### Option 1: Use the latest stable release
To install the latest stable release of this package
>```bash
>pip install --upgrade-strategy eager optimum[habana]
>```
The `--upgrade-strategy eager` option is needed to ensure `optimum-habana` is upgraded to the latest stable release.
To use the example associated with the latest stable release, run:
> ```
> git clone https://github.com/huggingface/optimum-habana
> cd optimum-habana && git checkout v1.11.1
> ```
> with `v1.11.1` the version number of this release.
### Option 2: Use the latest main branch under development
Optimum Habana is a fast-moving project, and you may want to install it from source and get the latest scripts :
```bash
pip install git+https://github.com/huggingface/optimum-habana.git
git clone https://github.com/huggingface/optimum-habana
```
## Install dependencies
To use DeepSpeed on HPUs, you also need to run the following command:
>```bash
>pip install git+https://github.com/HabanaAI/DeepSpeed.git@1.15.0
>```
To install the requirements for every example:
>```bash
>cd <example-folder>
>pip install -r requirements.txt
>```
## How to use it?
### Quick Start
🤗 Optimum Habana was designed with one goal in mind: **to make training and inference straightforward for any 🤗 Transformers and 🤗 Diffusers user while leveraging the complete power of Gaudi processors**.
#### Transformers Interface
There are two main classes one needs to know:
- [GaudiTrainer](https://huggingface.co/docs/optimum/habana/package_reference/trainer): the trainer class that takes care of compiling and distributing the model to run on HPUs, and performing training and evaluation.
- [GaudiConfig](https://huggingface.co/docs/optimum/habana/package_reference/gaudi_config): the class that enables to configure Habana Mixed Precision and to decide whether optimized operators and optimizers should be used or not.
The [GaudiTrainer](https://huggingface.co/docs/optimum/habana/package_reference/trainer) is very similar to the [🤗 Transformers Trainer](https://huggingface.co/docs/transformers/main_classes/trainer), and adapting a script using the Trainer to make it work with Gaudi will mostly consist in simply swapping the `Trainer` class for the `GaudiTrainer` one.
That's how most of the [example scripts](https://github.com/huggingface/optimum-habana/tree/main/examples) were adapted from their [original counterparts](https://github.com/huggingface/transformers/tree/main/examples/pytorch).
Here is an example:
```diff
- from transformers import Trainer, TrainingArguments
+ from optimum.habana import GaudiConfig, GaudiTrainer, GaudiTrainingArguments
- training_args = TrainingArguments(
+ training_args = GaudiTrainingArguments(
# training arguments...
+ use_habana=True,
+ use_lazy_mode=True, # whether to use lazy or eager mode
+ gaudi_config_name=path_to_gaudi_config,
)
# A lot of code here
# Initialize our Trainer
- trainer = Trainer(
+ trainer = GaudiTrainer(
model=model,
args=training_args, # Original training arguments.
train_dataset=train_dataset if training_args.do_train else None,
eval_dataset=eval_dataset if training_args.do_eval else None,
compute_metrics=compute_metrics,
tokenizer=tokenizer,
data_collator=data_collator,
)
```
where `gaudi_config_name` is the name of a model from the [Hub](https://huggingface.co/Habana) (Gaudi configurations are stored in model repositories) or a path to a local Gaudi configuration file (you can see [here](https://huggingface.co/docs/optimum/habana/package_reference/gaudi_config) how to write your own).
#### Diffusers Interface
You can generate images from prompts using Stable Diffusion on Gaudi using the [`GaudiStableDiffusionPipeline`](https://huggingface.co/docs/optimum/habana/package_reference/stable_diffusion_pipeline) class and the [`GaudiDDIMScheduler`] which have been both optimized for HPUs. Here is how to use them and the differences with the 🤗 Diffusers library:
```diff
- from diffusers import DDIMScheduler, StableDiffusionPipeline
+ from optimum.habana.diffusers import GaudiDDIMScheduler, GaudiStableDiffusionPipeline
model_name = "runwayml/stable-diffusion-v1-5"
- scheduler = DDIMScheduler.from_pretrained(model_name, subfolder="scheduler")
+ scheduler = GaudiDDIMScheduler.from_pretrained(model_name, subfolder="scheduler")
- pipeline = StableDiffusionPipeline.from_pretrained(
+ pipeline = GaudiStableDiffusionPipeline.from_pretrained(
model_name,
scheduler=scheduler,
+ use_habana=True,
+ use_hpu_graphs=True,
+ gaudi_config="Habana/stable-diffusion",
)
outputs = generator(
["An image of a squirrel in Picasso style"],
num_images_per_prompt=16,
+ batch_size=4,
)
```
### Documentation
Check out [the documentation of Optimum Habana](https://huggingface.co/docs/optimum/habana/index) for more advanced usage.
## Validated Models
The following model architectures, tasks and device distributions have been validated for 🤗 Optimum Habana:
> In the tables below, :heavy_check_mark: means single-card, multi-card and DeepSpeed have all been validated.
- Transformers:
<div align="center">
| Architecture | Training | Inference | <center>Tasks</center> |
|--------------|:--------:|:---------:|:-----------------------|
| BERT | :heavy_check_mark: | :heavy_check_mark: | <li>[text classification](https://github.com/huggingface/optimum-habana/tree/main/examples/text-classification)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
| RoBERTa | :heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
| ALBERT | :heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
| DistilBERT |:heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |
| GPT2 | :heavy_check_mark: | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| BLOOM(Z) | | <div style="text-align:left"><li>DeepSpeed</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| StarCoder | | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| GPT-J | <div style="text-align:left"><li>DeepSpeed</li></div> | <div style="text-align:left"><li>Single card</li><li>DeepSpeed</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| GPT-NeoX | <div style="text-align:left"><li>DeepSpeed</li></div> | <div style="text-align:left"><li>DeepSpeed</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| OPT | | <div style="text-align:left"><li>DeepSpeed</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| Llama 2 / CodeLlama / Llama 3 | :heavy_check_mark: | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| StableLM | | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| Falcon | <div style="text-align:left"><li>LoRA</li></div> | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| CodeGen | | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| MPT | | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| Mistral | | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| Phi | :heavy_check_mark: | <div style="text-align:left"><li>Single card</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| Mixtral | | <div style="text-align:left"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |
| T5 / Flan T5 | :heavy_check_mark: | :heavy_check_mark: | <li>[summarization](https://github.com/huggingface/optimum-habana/tree/main/examples/summarization)</li><li>[translation](https://github.com/huggingface/optimum-habana/tree/main/examples/translation)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering#fine-tuning-t5-on-squad20)</li> |
| BART | | <div style="text-align:left"><li>Single card</li></div> | <li>[summarization](https://github.com/huggingface/optimum-habana/tree/main/examples/summarization)</li><li>[translation](https://github.com/huggingface/optimum-habana/tree/main/examples/translation)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering#fine-tuning-t5-on-squad20)</li> |
| ViT | :heavy_check_mark: | :heavy_check_mark: | <li>[image classification](https://github.com/huggingface/optimum-habana/tree/main/examples/image-classification)</li> |
| Swin | :heavy_check_mark: | :heavy_check_mark: | <li>[image classification](https://github.com/huggingface/optimum-habana/tree/main/examples/image-classification)</li> |
| Wav2Vec2 | :heavy_check_mark: | :heavy_check_mark: | <li>[audio classification](https://github.com/huggingface/optimum-habana/tree/main/examples/audio-classification)</li><li>[speech recognition](https://github.com/huggingface/optimum-habana/tree/main/examples/speech-recognition)</li> |
| Whisper | :heavy_check_mark: | :heavy_check_mark: | <li>[speech recognition](https://github.com/huggingface/optimum-habana/tree/main/examples/speech-recognition)</li> |
| SpeechT5 | | <div style="text-align:left"><li>Single card</li></div> | <li>[text to speech](https://github.com/huggingface/optimum-habana/tree/main/examples/text-to-speech)</li> |
| CLIP | :heavy_check_mark: | :heavy_check_mark: | <li>[contrastive image-text training](https://github.com/huggingface/optimum-habana/tree/main/examples/contrastive-image-text)</li> |
| BridgeTower | :heavy_check_mark: | :heavy_check_mark: | <li>[contrastive image-text training](https://github.com/huggingface/optimum-habana/tree/main/examples/contrastive-image-text)</li> |
| ESMFold | | <div style="text-align:left"><li>Single card</li></div> | <li>[protein folding](https://github.com/huggingface/optimum-habana/tree/main/examples/protein-folding)</li> |
| Blip | | <div style="text-align:left"><li>Single card</li></div> | <li>[visual question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/visual-question-answering)</li><li>[image to text](https://github.com/huggingface/optimum-habana/tree/main/examples/image-to-text)</li> |
</div>
- Diffusers:
<div align="center">
| Architecture | Training | Inference | Tasks |
|------------------|:--------:|:--------------------:|:------|
| Stable Diffusion | | <li>Single card</li> | <li>[text-to-image generation](https://github.com/huggingface/optimum-habana/tree/main/examples/stable-diffusion)</li> |
| LDM3D | | <li>Single card</li> | <li>[text-to-image generation](https://github.com/huggingface/optimum-habana/tree/main/examples/stable-diffusion)</li> |
</div>
- TRL:
<div align="center">
| Architecture | Training | Inference | Tasks |
|------------------|:--------:|:--------------------:|:------|
| Llama 2 | :heavy_check_mark: | | <li>[DPO Pipeline](https://github.com/huggingface/optimum-habana/tree/main/examples/trl)</li> |
| Llama 2 | :heavy_check_mark: | | <li>[PPO Pipeline](https://github.com/huggingface/optimum-habana/tree/main/examples/trl)</li> |
</div>
Other models and tasks supported by the 🤗 Transformers and 🤗 Diffusers library may also work. You can refer to this [section](https://github.com/huggingface/optimum-habana#how-to-use-it) for using them with 🤗 Optimum Habana. Besides, [this page](https://github.com/huggingface/optimum-habana/tree/main/examples) explains how to modify any [example](https://github.com/huggingface/transformers/tree/main/examples/pytorch) from the 🤗 Transformers library to make it work with 🤗 Optimum Habana.
If you find any issues while using those, please open an issue or a pull request.
## Gaudi Setup
Please refer to Habana Gaudi's official [installation guide](https://docs.habana.ai/en/latest/Installation_Guide/index.html).
> Tests should be run in a Docker container based on Habana Docker images.
>
> The current version has been validated for SynapseAI 1.15.
## Development
Check the [contributor guide](https://github.com/huggingface/optimum/blob/main/CONTRIBUTING.md) for instructions.
Raw data
{
"_id": null,
"home_page": "https://huggingface.co/hardware/habana",
"name": "optimum-habana",
"maintainer": null,
"docs_url": null,
"requires_python": null,
"maintainer_email": null,
"keywords": "transformers, diffusers, mixed-precision training, fine-tuning, gaudi, hpu",
"author": "HuggingFace Inc. Special Ops Team",
"author_email": "hardware@huggingface.co",
"download_url": "https://files.pythonhosted.org/packages/21/a8/f5fa85975b602ef568d9869f3fb594d282e722f6213e42f86e082f0fb963/optimum_habana-1.11.1.tar.gz",
"platform": null,
"description": "<!---\nCopyright 2022 The HuggingFace Team. All rights reserved.\n\nLicensed under the Apache License, Version 2.0 (the \"License\");\nyou may not use this file except in compliance with the License.\nYou may obtain a copy of the License at\n\n http://www.apache.org/licenses/LICENSE-2.0\n\nUnless required by applicable law or agreed to in writing, software\ndistributed under the License is distributed on an \"AS IS\" BASIS,\nWITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\nSee the License for the specific language governing permissions and\nlimitations under the License.\n-->\n\n![](https://github.com/huggingface/optimum-habana/blob/main/readme_logo.png)\n\n\n# Optimum Habana\n\n\ud83e\udd17 Optimum Habana is the interface between the \ud83e\udd17 Transformers and Diffusers libraries and [Habana's Gaudi processor (HPU)](https://docs.habana.ai/en/latest/index.html).\nIt provides a set of tools enabling easy model loading, training and inference on single- and multi-HPU settings for different downstream tasks.\nThe list of officially validated models and tasks is available [here](https://github.com/huggingface/optimum-habana#validated-models). Users can try other models and tasks with only few changes.\n\n\n## What is a Habana Processing Unit (HPU)?\n\nHPUs offer fast model training and inference as well as a great price-performance ratio.\nCheck out [this blog post about BERT pre-training](https://huggingface.co/blog/pretraining-bert) and [this article benchmarking Habana Gaudi2 versus Nvidia A100 GPUs](https://huggingface.co/blog/habana-gaudi-2-benchmark) for concrete examples.\nIf you are not familiar with HPUs and would like to know more about them, we recommend you take a look at [our conceptual guide](https://huggingface.co/docs/optimum/habana/concept_guides/hpu).\n\n\n## Install the library and get example scripts\n\n### Option 1: Use the latest stable release\n\nTo install the latest stable release of this package\n>```bash\n>pip install --upgrade-strategy eager optimum[habana]\n>```\n\nThe `--upgrade-strategy eager` option is needed to ensure `optimum-habana` is upgraded to the latest stable release.\n\nTo use the example associated with the latest stable release, run:\n> ```\n> git clone https://github.com/huggingface/optimum-habana\n> cd optimum-habana && git checkout v1.11.1\n> ```\n> with `v1.11.1` the version number of this release.\n\n### Option 2: Use the latest main branch under development\n\nOptimum Habana is a fast-moving project, and you may want to install it from source and get the latest scripts :\n\n```bash\npip install git+https://github.com/huggingface/optimum-habana.git\ngit clone https://github.com/huggingface/optimum-habana\n```\n\n## Install dependencies\n\nTo use DeepSpeed on HPUs, you also need to run the following command:\n>```bash\n>pip install git+https://github.com/HabanaAI/DeepSpeed.git@1.15.0\n>```\n\nTo install the requirements for every example:\n>```bash\n>cd <example-folder>\n>pip install -r requirements.txt\n>```\n\n\n## How to use it?\n\n### Quick Start\n\n\ud83e\udd17 Optimum Habana was designed with one goal in mind: **to make training and inference straightforward for any \ud83e\udd17 Transformers and \ud83e\udd17 Diffusers user while leveraging the complete power of Gaudi processors**.\n\n#### Transformers Interface\n\nThere are two main classes one needs to know:\n- [GaudiTrainer](https://huggingface.co/docs/optimum/habana/package_reference/trainer): the trainer class that takes care of compiling and distributing the model to run on HPUs, and performing training and evaluation.\n- [GaudiConfig](https://huggingface.co/docs/optimum/habana/package_reference/gaudi_config): the class that enables to configure Habana Mixed Precision and to decide whether optimized operators and optimizers should be used or not.\n\nThe [GaudiTrainer](https://huggingface.co/docs/optimum/habana/package_reference/trainer) is very similar to the [\ud83e\udd17 Transformers Trainer](https://huggingface.co/docs/transformers/main_classes/trainer), and adapting a script using the Trainer to make it work with Gaudi will mostly consist in simply swapping the `Trainer` class for the `GaudiTrainer` one.\nThat's how most of the [example scripts](https://github.com/huggingface/optimum-habana/tree/main/examples) were adapted from their [original counterparts](https://github.com/huggingface/transformers/tree/main/examples/pytorch).\n\nHere is an example:\n```diff\n- from transformers import Trainer, TrainingArguments\n+ from optimum.habana import GaudiConfig, GaudiTrainer, GaudiTrainingArguments\n\n- training_args = TrainingArguments(\n+ training_args = GaudiTrainingArguments(\n # training arguments...\n+ use_habana=True,\n+ use_lazy_mode=True, # whether to use lazy or eager mode\n+ gaudi_config_name=path_to_gaudi_config,\n)\n\n# A lot of code here\n\n# Initialize our Trainer\n- trainer = Trainer(\n+ trainer = GaudiTrainer(\n model=model,\n args=training_args, # Original training arguments.\n train_dataset=train_dataset if training_args.do_train else None,\n eval_dataset=eval_dataset if training_args.do_eval else None,\n compute_metrics=compute_metrics,\n tokenizer=tokenizer,\n data_collator=data_collator,\n)\n```\n\nwhere `gaudi_config_name` is the name of a model from the [Hub](https://huggingface.co/Habana) (Gaudi configurations are stored in model repositories) or a path to a local Gaudi configuration file (you can see [here](https://huggingface.co/docs/optimum/habana/package_reference/gaudi_config) how to write your own).\n\n\n#### Diffusers Interface\n\nYou can generate images from prompts using Stable Diffusion on Gaudi using the [`GaudiStableDiffusionPipeline`](https://huggingface.co/docs/optimum/habana/package_reference/stable_diffusion_pipeline) class and the [`GaudiDDIMScheduler`] which have been both optimized for HPUs. Here is how to use them and the differences with the \ud83e\udd17 Diffusers library:\n\n```diff\n- from diffusers import DDIMScheduler, StableDiffusionPipeline\n+ from optimum.habana.diffusers import GaudiDDIMScheduler, GaudiStableDiffusionPipeline\n\n\nmodel_name = \"runwayml/stable-diffusion-v1-5\"\n\n- scheduler = DDIMScheduler.from_pretrained(model_name, subfolder=\"scheduler\")\n+ scheduler = GaudiDDIMScheduler.from_pretrained(model_name, subfolder=\"scheduler\")\n\n- pipeline = StableDiffusionPipeline.from_pretrained(\n+ pipeline = GaudiStableDiffusionPipeline.from_pretrained(\n model_name,\n scheduler=scheduler,\n+ use_habana=True,\n+ use_hpu_graphs=True,\n+ gaudi_config=\"Habana/stable-diffusion\",\n)\n\noutputs = generator(\n [\"An image of a squirrel in Picasso style\"],\n num_images_per_prompt=16,\n+ batch_size=4,\n)\n```\n\n\n### Documentation\n\nCheck out [the documentation of Optimum Habana](https://huggingface.co/docs/optimum/habana/index) for more advanced usage.\n\n\n## Validated Models\n\nThe following model architectures, tasks and device distributions have been validated for \ud83e\udd17 Optimum Habana:\n\n> In the tables below, :heavy_check_mark: means single-card, multi-card and DeepSpeed have all been validated.\n\n- Transformers:\n<div align=\"center\">\n\n| Architecture | Training | Inference | <center>Tasks</center> |\n|--------------|:--------:|:---------:|:-----------------------|\n| BERT | :heavy_check_mark: | :heavy_check_mark: | <li>[text classification](https://github.com/huggingface/optimum-habana/tree/main/examples/text-classification)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |\n| RoBERTa | :heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |\n| ALBERT | :heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |\n| DistilBERT |:heavy_check_mark: | :heavy_check_mark: | <li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering)</li><li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li> |\n| GPT2 | :heavy_check_mark: | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| BLOOM(Z) | | <div style=\"text-align:left\"><li>DeepSpeed</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| StarCoder | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| GPT-J | <div style=\"text-align:left\"><li>DeepSpeed</li></div> | <div style=\"text-align:left\"><li>Single card</li><li>DeepSpeed</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| GPT-NeoX | <div style=\"text-align:left\"><li>DeepSpeed</li></div> | <div style=\"text-align:left\"><li>DeepSpeed</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| OPT | | <div style=\"text-align:left\"><li>DeepSpeed</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| Llama 2 / CodeLlama / Llama 3 | :heavy_check_mark: | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| StableLM | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| Falcon | <div style=\"text-align:left\"><li>LoRA</li></div> | :heavy_check_mark: | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| CodeGen | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| MPT | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| Mistral | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| Phi | :heavy_check_mark: | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[language modeling](https://github.com/huggingface/optimum-habana/tree/main/examples/language-modeling)</li><li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| Mixtral | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[text generation](https://github.com/huggingface/optimum-habana/tree/main/examples/text-generation)</li> |\n| T5 / Flan T5 | :heavy_check_mark: | :heavy_check_mark: | <li>[summarization](https://github.com/huggingface/optimum-habana/tree/main/examples/summarization)</li><li>[translation](https://github.com/huggingface/optimum-habana/tree/main/examples/translation)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering#fine-tuning-t5-on-squad20)</li> |\n| BART | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[summarization](https://github.com/huggingface/optimum-habana/tree/main/examples/summarization)</li><li>[translation](https://github.com/huggingface/optimum-habana/tree/main/examples/translation)</li><li>[question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/question-answering#fine-tuning-t5-on-squad20)</li> |\n| ViT | :heavy_check_mark: | :heavy_check_mark: | <li>[image classification](https://github.com/huggingface/optimum-habana/tree/main/examples/image-classification)</li> |\n| Swin | :heavy_check_mark: | :heavy_check_mark: | <li>[image classification](https://github.com/huggingface/optimum-habana/tree/main/examples/image-classification)</li> |\n| Wav2Vec2 | :heavy_check_mark: | :heavy_check_mark: | <li>[audio classification](https://github.com/huggingface/optimum-habana/tree/main/examples/audio-classification)</li><li>[speech recognition](https://github.com/huggingface/optimum-habana/tree/main/examples/speech-recognition)</li> |\n| Whisper | :heavy_check_mark: | :heavy_check_mark: | <li>[speech recognition](https://github.com/huggingface/optimum-habana/tree/main/examples/speech-recognition)</li> |\n| SpeechT5 | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[text to speech](https://github.com/huggingface/optimum-habana/tree/main/examples/text-to-speech)</li> |\n| CLIP | :heavy_check_mark: | :heavy_check_mark: | <li>[contrastive image-text training](https://github.com/huggingface/optimum-habana/tree/main/examples/contrastive-image-text)</li> |\n| BridgeTower | :heavy_check_mark: | :heavy_check_mark: | <li>[contrastive image-text training](https://github.com/huggingface/optimum-habana/tree/main/examples/contrastive-image-text)</li> |\n| ESMFold | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[protein folding](https://github.com/huggingface/optimum-habana/tree/main/examples/protein-folding)</li> |\n| Blip | | <div style=\"text-align:left\"><li>Single card</li></div> | <li>[visual question answering](https://github.com/huggingface/optimum-habana/tree/main/examples/visual-question-answering)</li><li>[image to text](https://github.com/huggingface/optimum-habana/tree/main/examples/image-to-text)</li> |\n\n</div>\n\n- Diffusers:\n\n<div align=\"center\">\n\n| Architecture | Training | Inference | Tasks |\n|------------------|:--------:|:--------------------:|:------|\n| Stable Diffusion | | <li>Single card</li> | <li>[text-to-image generation](https://github.com/huggingface/optimum-habana/tree/main/examples/stable-diffusion)</li> |\n| LDM3D | | <li>Single card</li> | <li>[text-to-image generation](https://github.com/huggingface/optimum-habana/tree/main/examples/stable-diffusion)</li> |\n\n</div>\n\n- TRL:\n\n<div align=\"center\">\n\n| Architecture | Training | Inference | Tasks |\n|------------------|:--------:|:--------------------:|:------|\n| Llama 2 | :heavy_check_mark: | | <li>[DPO Pipeline](https://github.com/huggingface/optimum-habana/tree/main/examples/trl)</li> |\n| Llama 2 | :heavy_check_mark: | | <li>[PPO Pipeline](https://github.com/huggingface/optimum-habana/tree/main/examples/trl)</li> |\n\n</div>\n\nOther models and tasks supported by the \ud83e\udd17 Transformers and \ud83e\udd17 Diffusers library may also work. You can refer to this [section](https://github.com/huggingface/optimum-habana#how-to-use-it) for using them with \ud83e\udd17 Optimum Habana. Besides, [this page](https://github.com/huggingface/optimum-habana/tree/main/examples) explains how to modify any [example](https://github.com/huggingface/transformers/tree/main/examples/pytorch) from the \ud83e\udd17 Transformers library to make it work with \ud83e\udd17 Optimum Habana.\n\nIf you find any issues while using those, please open an issue or a pull request.\n\n\n## Gaudi Setup\n\nPlease refer to Habana Gaudi's official [installation guide](https://docs.habana.ai/en/latest/Installation_Guide/index.html).\n\n> Tests should be run in a Docker container based on Habana Docker images.\n>\n> The current version has been validated for SynapseAI 1.15.\n\n\n## Development\n\nCheck the [contributor guide](https://github.com/huggingface/optimum/blob/main/CONTRIBUTING.md) for instructions.\n",
"bugtrack_url": null,
"license": "Apache",
"summary": "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 model loading, training and inference on single- and multi-HPU settings for different downstream tasks.",
"version": "1.11.1",
"project_urls": {
"Homepage": "https://huggingface.co/hardware/habana"
},
"split_keywords": [
"transformers",
" diffusers",
" mixed-precision training",
" fine-tuning",
" gaudi",
" hpu"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "0f4c754da8454d30839e9143d803e41d31604bd8196a0c6ba49ee306195d1266",
"md5": "44aa7539f789d035a7ca334e0a8a3383",
"sha256": "2b4f2a801ebd7f235bde91f3e351893efcc586afef3b0c867eb3c4c8c4c26bec"
},
"downloads": -1,
"filename": "optimum_habana-1.11.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "44aa7539f789d035a7ca334e0a8a3383",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 347456,
"upload_time": "2024-04-20T05:15:47",
"upload_time_iso_8601": "2024-04-20T05:15:47.144509Z",
"url": "https://files.pythonhosted.org/packages/0f/4c/754da8454d30839e9143d803e41d31604bd8196a0c6ba49ee306195d1266/optimum_habana-1.11.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "21a8f5fa85975b602ef568d9869f3fb594d282e722f6213e42f86e082f0fb963",
"md5": "2b4efea56a54e903d42f51d7669c4ca3",
"sha256": "c70dbd6d69ce11c8bb78292913304616600d051d94237c8310c3bdbf73ca7017"
},
"downloads": -1,
"filename": "optimum_habana-1.11.1.tar.gz",
"has_sig": false,
"md5_digest": "2b4efea56a54e903d42f51d7669c4ca3",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 313449,
"upload_time": "2024-04-20T05:15:49",
"upload_time_iso_8601": "2024-04-20T05:15:49.642202Z",
"url": "https://files.pythonhosted.org/packages/21/a8/f5fa85975b602ef568d9869f3fb594d282e722f6213e42f86e082f0fb963/optimum_habana-1.11.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-04-20 05:15:49",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "optimum-habana"
}