Terjemahan disediakan oleh mesin penerjemah. Jika konten terjemahan yang diberikan bertentangan dengan versi bahasa Inggris aslinya, utamakan versi bahasa Inggris.

# Kerangka Kerja yang Didukung, Wilayah AWS, Jenis Instans, dan Model yang Diuji
<a name="training-compiler-support"></a>

**penting**  
Amazon Web Services (AWS) mengumumkan bahwa tidak akan ada rilis baru atau versi Tra SageMaker ining Compiler. Anda dapat terus menggunakan Tra SageMaker ining Compiler melalui AWS Deep Learning Containers (DLC) yang ada untuk SageMaker Pelatihan. Penting untuk dicatat bahwa sementara DLC yang ada tetap dapat diakses, mereka tidak akan lagi menerima tambalan atau pembaruan dari AWS, sesuai dengan Kebijakan [ Dukungan Kerangka Kon ](https://docs.aws.amazon.com/deep-learning-containers/latest/devguide/support-policy.html)AWS tainer Pembelajaran Mendalam.

Sebelum menggunakan SageMaker Training Compiler, periksa apakah kerangka kerja pilihan Anda didukung, jenis instans tersedia di AWS akun Anda, dan AWS akun Anda ada di salah satu yang didukung Wilayah AWS.

**catatan**  
SageMaker Training Compiler tersedia di SageMaker Python SDK v2.70.0 atau yang lebih baru.

## Kerangka Kerja yang Didukung
<a name="training-compiler-supported-frameworks"></a>

SageMaker Training Compiler mendukung kerangka kerja pembelajaran mendalam berikut dan tersedia melalui AWS Deep Learning Containers.

**Topics**
+ [PyTorch](#training-compiler-supported-frameworks-pytorch)
+ [TensorFlow](#training-compiler-supported-frameworks-tensorflow)

### PyTorch
<a name="training-compiler-supported-frameworks-pytorch"></a>



- **PyTorch**
  - **Versi kerangka kerja:** PyTorch v1.13.1 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/pytorch-trcomp-training: 1.12.0-gpu-py38-cu113-ubuntu20.04-sagemaker / **Dapat diperpanjang untuk kustomisasi Docker:** Tidak
  - **Versi kerangka kerja:** PyTorch v1.12.0 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/pytorch-trcomp-training: 1.13.1-gpu-py39-cu117-ubuntu20.04-sagemaker / **Dapat diperpanjang untuk kustomisasi Docker:** Tidak

- **PyTorch dengan Hugging Face Transformers**
  - **Versi kerangka kerja:** Transformator v4.21.1<br />PyTorch v1.11.0 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/huggingface-pytorch-trcomp-pelatihan: 1.11.0-transformers4.21.1-gpu-py38-cu113-ubuntu20.04 / **Dapat diperpanjang untuk kustomisasi Docker:** Tidak
  - **Versi kerangka kerja:** Transformer v4.17.0<br />PyTorch v1.10.2 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/huggingface-pytorch-trcomp-pelatihan: 1.10.2-transformers4.17.0-gpu-py38-cu113-ubuntu20.04 / **Dapat diperpanjang untuk kustomisasi Docker:** Tidak
  - **Versi kerangka kerja:** Transformator v4.11.0<br />PyTorch v1.9.0 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/huggingface-pytorch-training-comp:1.9.0-transformers4.11.0-gpu-py38-cu111-ubuntu20.04 / **Dapat diperpanjang untuk kustomisasi Docker:** Tidak



### TensorFlow
<a name="training-compiler-supported-frameworks-tensorflow"></a>



- **TensorFlow**
  - **Versi kerangka kerja:** TensorFlow v2.11.0 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/tensorflow-pelatihan: 2.11.0-gpu-py39-cu112-ubuntu20.04-sagemaker / **Dapat diperpanjang untuk kustomisasi Docker:** Ya
  - **Versi kerangka kerja:** TensorFlow v2.10.0 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/tensorflow-pelatihan: 2.10.0-gpu-py39-cu112-ubuntu20.04-sagemaker / **Dapat diperpanjang untuk kustomisasi Docker:** Ya
  - **Versi kerangka kerja:** TensorFlow v2.9.1 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/tensorflow-pelatihan: 2.9.1-gpu-py39-cu112-ubuntu20.04-sagemaker / **Dapat diperpanjang untuk kustomisasi Docker:** Ya

- **TensorFlow dengan Hugging Face Transformers**
  - **Versi kerangka kerja:** Transformer v4.17.0<br />TensorFlow v2.6.3 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/huggingface-tensorflow-trcomp-training: 2.6.3-transformers4.17.0-gpu-py38-cu112-ubuntu20.04 / **Dapat diperpanjang untuk kustomisasi Docker:** Tidak
  - **Versi kerangka kerja:** Transformator v4.11.0<br />TensorFlow v2.5.1 / **URI Wadah Pembelajaran Mendalam:** 763104351884.dkr.ep. {{<region>}}.amazonaw. com/huggingface-tensorflow-training-comp:2.5.1-transformers4.11.0-gpu-py37-cu112-ubuntu18.04 / **Dapat diperpanjang untuk kustomisasi Docker:** Tidak



Untuk informasi selengkapnya, lihat Gambar [ yang Tersedia ](https://github.com/aws/deep-learning-containers/blob/master/available_images.md) di GitHub repositori *AWS Deep Learning Containers*.

## Wilayah AWS
<a name="training-compiler-availablity-zone"></a>

Kon [ SageMaker tainer Kompil ](https://github.com/aws/deep-learning-containers/blob/master/available_images.md#sagemaker-training-compiler-containers) er Pelatihan tersedia di Wilayah AWS tempat [AWS Deep Learning ](https://github.com/aws/deep-learning-containers/blob/master/available_images.md) Container beroperasi kecuali wilayah China.

## Tipe Instans Yang Didukung
<a name="training-compiler-supported-instance-types"></a>

SageMaker Training Compiler diuji dan mendukung jenis instance ML berikut.
+ Instans P4
+ Instans P3
+ Instans G4dn
+ Instans G5

Untuk spesifikasi jenis instans, lihat ** bagian Komput ** asi Dipercepat di halaman [ Jenis Instans ](https://aws.amazon.com/ec2/instance-types/) Amazon EC2. Untuk informasi tentang harga instans, lihat [ SageMaker Harga Amazon](https://aws.amazon.com/sagemaker/pricing/).

Jika Anda menemukan pesan kesalahan yang mirip dengan berikut ini, ikuti petunjuk di [ Minta peningkatan kuota layanan untuk sumber daya SageMaker AI](https://docs.aws.amazon.com/sagemaker/latest/dg/regions-quotas.html#service-limit-increase-request-procedure).

```
ResourceLimitExceeded: An error occurred (ResourceLimitExceeded) when calling
the CreateTrainingJob operation: The account-level service limit 'ml.p3dn.24xlarge
for training job usage' is 0 Instances, with current utilization of 0 Instances
and a request delta of 1 Instances.
Please contact AWS support to request an increase for this limit.
```

## Model yang Diuji
<a name="training-compiler-tested-models"></a>

Tabel berikut mencakup daftar model yang telah diuji dengan Tra SageMaker ining Compiler. Sebagai referensi, ukuran batch terbesar yang dapat masuk ke dalam memori juga disertakan bersama parameter pelatihan lainnya. SageMaker Training Compiler dapat mengubah jejak memori dari proses pelatihan model; sebagai hasilnya, ukuran batch yang lebih besar sering dapat digunakan selama proses pelatihan, yang selanjutnya mengurangi total waktu pelatihan. Dalam beberapa kasus, SageMaker Training Compiler secara cerdas mempromosikan caching yang mengarah pada penurunan ukuran batch terbesar yang dapat muat pada GPU. Anda harus menyetel ulang hyperparameter model Anda dan menemukan ukuran batch yang optimal untuk kasus Anda. Untuk menghemat waktu, gunakan tabel referensi berikut untuk mencari ukuran batch yang dapat menjadi titik awal yang baik untuk kasus penggunaan Anda.

**catatan**  
Ukuran batch adalah ukuran batch lokal yang sesuai dengan masing-masing GPU dalam jenis instance masing-masing. Anda juga harus menyesuaikan tingkat pembelajaran saat mengubah ukuran batch.

### PyTorch 1.13.1
<a name="training-compiler-tested-models-pt1131"></a>

**Model pemrosesan bahasa alami (NLP) **

Model berikut diuji untuk pekerjaan pelatihan untuk semua kombinasi node tunggal dan multi-node dengan inti GPU tunggal atau multi dan Precision Campuran Otomatis (AMP) seperti yang ditunjukkan.


<table>
<thead>
  <tr><th colspan="7">Single-node/multi-nodetunggal- GPU/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Presisi</th><th>Panjang Urutan</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>80</td><td>192</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>128</td><td>332</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>80</td><td>224</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>160</td><td>288</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>160</td><td>280</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>240</td><td>472</td></tr>
  <tr><td>distilgpt2</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>77</td><td>128</td></tr>
  <tr><td>distilgpt2</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>138</td><td>390</td></tr>
  <tr><td>distilgpt2</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>96</td><td>256</td></tr>
  <tr><td>distilroberta-base</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>96</td><td>192</td></tr>
  <tr><td>distilroberta-base</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>171</td><td>380</td></tr>
  <tr><td>distilroberta-base</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>112</td><td>256</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>52</td><td>152</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>84</td><td>240</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>58</td><td>164</td></tr>
  <tr><td>microsoft/deberta-basis</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>48</td><td>128</td></tr>
  <tr><td>microsoft/deberta-basis</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>84</td><td>207</td></tr>
  <tr><td>microsoft/deberta-basis</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>53</td><td>133</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>125</td><td>224</td></tr>
  <tr><td>xlm-roberta-base</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>16</td><td>31</td></tr>
  <tr><td>xlm-roberta-base</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>18</td><td>50</td></tr>
  <tr><td>xlnet-base casing</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>128</td><td>240</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-103-v1</td><td>g5.48xlarge</td><td>mengapung16</td><td>512</td><td>29</td><td>50</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-103-v1</td><td>g5.48xlarge</td><td>mengapung16</td><td>512</td><td>45</td><td>64</td></tr>
  <tr><td>gpt2</td><td>wikitext-103-v1</td><td>g5.48xlarge</td><td>mengapung16</td><td>512</td><td>18</td><td>45</td></tr>
  <tr><td>Roberta base</td><td>wikitext-103-v1</td><td>g5.48xlarge</td><td>mengapung16</td><td>512</td><td>23</td><td>44</td></tr>
  <tr><td>gpt2</td><td>wikitext-103-v1</td><td>p4d.24xlarge</td><td>mengapung16</td><td>512</td><td>36</td><td>64</td></tr>
</tbody>
</table>


**Model Visi Komputer (CV) **

Diuji menggunakan [ TensorFlow Model Garden ](https://github.com/tensorflow/models) dengan Automatic Mixed Precision (AMP) seperti yang ditunjukkan.


<table>
<thead>
  <tr><th colspan="6">Single/multi-node single/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Presisi</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>ResNet152</td><td>makanan101</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>144</td></tr>
  <tr><td>ResNet152</td><td>makanan101</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>192</td></tr>
  <tr><td>ResNet152</td><td>makanan101</td><td>p3.2xlarge</td><td>mengapung16</td><td>152</td><td>156</td></tr>
  <tr><td>ViT</td><td>makanan101</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>512</td><td>512</td></tr>
  <tr><td>ViT</td><td>makanan101</td><td>g5.4xlarge</td><td>mengapung16</td><td>992</td><td>768</td></tr>
  <tr><td>ViT</td><td>makanan101</td><td>p3.2xlarge</td><td>mengapung16</td><td>848</td><td>768</td></tr>
</tbody>
</table>


### PyTorch 1.12.0
<a name="training-compiler-tested-models-pt1120"></a>

**Model pemrosesan bahasa alami (NLP) **

Model berikut diuji untuk pekerjaan pelatihan untuk semua kombinasi node tunggal dan multi-node dengan inti GPU tunggal atau multi dan Precision Campuran Otomatis (AMP) seperti yang ditunjukkan.


<table>
<thead>
  <tr><th colspan="7">Single-node/multi-nodetunggal- GPU/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Presisi</th><th>Panjang Urutan</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>128</td><td>248</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>160</td><td>288</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>160</td><td>279</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>105</td><td>164</td></tr>
  <tr><td>distilgpt2</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>136</td><td>256</td></tr>
  <tr><td>distilgpt2</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>80</td><td>118</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>84</td><td>240</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>80</td><td>119</td></tr>
  <tr><td>microsoft/deberta-basis</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>93</td><td>197</td></tr>
  <tr><td>microsoft/deberta-basis</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>113</td><td>130</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>125</td><td>224</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>78</td><td>112</td></tr>
  <tr><td>xlnet-base casing</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>138</td><td>240</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>mengapung16</td><td>512</td><td></td><td>52</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>mengapung16</td><td>512</td><td></td><td>160</td></tr>
  <tr><td>gpt2</td><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>mengapung16</td><td>512</td><td></td><td>25</td></tr>
  <tr><td>Roberta base</td><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>mengapung16</td><td>512</td><td></td><td>64</td></tr>
</tbody>
</table>


### TensorFlow 2.11.0
<a name="training-compiler-tested-models-tf2110"></a>

**Model Visi Komputer (CV) **

Diuji menggunakan [ TensorFlow Model Garden ](https://github.com/tensorflow/models) dengan Automatic Mixed Precision (AMP) seperti yang ditunjukkan.


<table>
<thead>
  <tr><th colspan="6">Single/multi-node single/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Presisi</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>MaskRCNN-ResNet50-FPN</td><td>COCO-2017</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>6</td><td>8</td></tr>
  <tr><td>MaskRCNN-ResNet50-FPN</td><td>COCO-2017</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>4</td><td>6</td></tr>
  <tr><td>ResNet50</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>192</td><td>256</td></tr>
  <tr><td>ResNet50</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>256</td><td>256</td></tr>
  <tr><td>ResNet101</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>256</td></tr>
  <tr><td>ResNet101</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>128</td></tr>
  <tr><td>ResNet152</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>224</td></tr>
  <tr><td>ResNet152</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>128</td></tr>
  <tr><td>VisionTransformer</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>112</td><td>144</td></tr>
  <tr><td>VisionTransformer</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>96</td><td>128</td></tr>
</tbody>
</table>


**Model Pemrosesan Bahasa Alami (NLP) **

Diuji menggunakan model [ Transformer ](https://github.com/huggingface/transformers) dengan `Sequence_Len=128` dan Automatic Mixed Precision (AMP) seperti yang ditunjukkan.


<table>
<thead>
  <tr><th colspan="6">Single/multi-node single/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Presisi</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>160</td><td>197</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>95</td><td>127</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>160</td><td>128</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>104</td><td>111</td></tr>
  <tr><td>bert-large-uncased</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>65</td><td>48</td></tr>
  <tr><td>bert-large-uncased</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>40</td><td>35</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>162</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>105</td><td>111</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>256</td><td>264</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>169</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>120</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>80</td><td>83</td></tr>
  <tr><td>jplu/tf-xlm-roberta-basis</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>32</td><td>32</td></tr>
  <tr><td>jplu/tf-xlm-roberta-basis</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>32</td><td>36</td></tr>
  <tr><td>microsoft/mpnet-basis</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>144</td><td>160</td></tr>
  <tr><td>microsoft/mpnet-basis</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>106</td><td>110</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>128</td><td>128</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>72</td><td>98</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>128</td><td>192</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>ml.p3.16xlarge</td><td>mengapung16</td><td>95</td><td>96</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>256</td><td>256</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>ml.p3.16xlarge</td><td>mengapung16</td><td>140</td><td>184</td></tr>
  <tr><td>google/electra-diskriminator kecil</td><td>wikitext-2-raw-v1</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>256</td><td>384</td></tr>
  <tr><td>google/electra-diskriminator kecil</td><td>wikitext-2-raw-v1</td><td>ml.p3.16xlarge</td><td>mengapung16</td><td>256</td><td>268</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>116</td><td>116</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>ml.p3.16xlarge</td><td>mengapung16</td><td>85</td><td>83</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>ml.p4d.24xbesar</td><td>mengapung16</td><td>94</td><td>110</td></tr>
  <tr><td>microsoft/mpnet-basis</td><td>wikitext-2-raw-v1</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>187</td><td>164</td></tr>
  <tr><td>microsoft/mpnet-basis</td><td>wikitext-2-raw-v1</td><td>ml.p3.16xlarge</td><td>mengapung16</td><td>106</td><td>111</td></tr>
</tbody>
</table>


### TensorFlow 2.0.0
<a name="training-compiler-tested-models-tf2100"></a>

**Model Visi Komputer (CV) **

Diuji menggunakan [ TensorFlow Model Garden ](https://github.com/tensorflow/models) dengan Automatic Mixed Precision (AMP) seperti yang ditunjukkan.


<table>
<thead>
  <tr><th colspan="6">Single-nodetunggal- GPU/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Presisi</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>DetectionTransformer-ResNet50</td><td>COCO-2017</td><td>ml.g4dn.2xbesar</td><td>mengapung32</td><td>2</td><td>4</td></tr>
  <tr><td>DetectionTransformer-ResNet50</td><td>COCO-2017</td><td>ml.g5.2xbesar</td><td>mengapung32</td><td>3</td><td>6</td></tr>
  <tr><td>DetectionTransformer-ResNet50</td><td>COCO-2017</td><td>ml.p3.2xlarge</td><td>mengapung32</td><td>2</td><td>4</td></tr>
  <tr><td>MaskRCNN-ResNet50-FPN</td><td>COCO-2017</td><td>ml.g4dn.2xbesar</td><td>mengapung16</td><td>4</td><td>6</td></tr>
  <tr><td>MaskRCNN-ResNet50-FPN</td><td>COCO-2017</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>6</td><td>8</td></tr>
  <tr><td>MaskRCNN-ResNet50-FPN</td><td>COCO-2017</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>48</td><td>64</td></tr>
  <tr><td>MaskRCNN-ResNet50-FPN</td><td>COCO-2017</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>4</td><td>6</td></tr>
  <tr><td>ResNet50</td><td>ImageNet</td><td>ml.g4dn.2xbesar</td><td>mengapung16</td><td>224</td><td>256</td></tr>
  <tr><td>ResNet50</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>192</td><td>160</td></tr>
  <tr><td>ResNet50</td><td>ImageNet</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>2048</td><td>2048</td></tr>
  <tr><td>ResNet50</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>224</td><td>160</td></tr>
  <tr><td>ResNet101</td><td>ImageNet</td><td>ml.g4dn.2xbesar</td><td>mengapung16</td><td>160</td><td>128</td></tr>
  <tr><td>ResNet101</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>192</td><td>256</td></tr>
  <tr><td>ResNet101</td><td>ImageNet</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>2048</td><td>2048</td></tr>
  <tr><td>ResNet101</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>160</td><td>224</td></tr>
  <tr><td>ResNet152</td><td>ImageNet</td><td>ml.g4dn.2xbesar</td><td>mengapung16</td><td>128</td><td>128</td></tr>
  <tr><td>ResNet152</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>192</td><td>224</td></tr>
  <tr><td>ResNet152</td><td>ImageNet</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>1536</td><td>1792</td></tr>
  <tr><td>ResNet152</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>128</td><td>160</td></tr>
  <tr><td>VisionTransformer</td><td>ImageNet</td><td>ml.g4dn.2xbesar</td><td>mengapung16</td><td>80</td><td>128</td></tr>
  <tr><td>VisionTransformer</td><td>ImageNet</td><td>ml.g5.2xbesar</td><td>mengapung16</td><td>112</td><td>144</td></tr>
  <tr><td>VisionTransformer</td><td>ImageNet</td><td>ml.g5.48xbesar</td><td>mengapung16</td><td>896</td><td>1152</td></tr>
  <tr><td>VisionTransformer</td><td>ImageNet</td><td>ml.p3.2xlarge</td><td>mengapung16</td><td>80</td><td>128</td></tr>
</tbody>
</table>


**Model Pemrosesan Bahasa Alami (NLP) **

Diuji menggunakan model [ Transformer ](https://github.com/huggingface/transformers) dengan `Sequence_Len=128` dan Automatic Mixed Precision (AMP) seperti yang ditunjukkan.


<table>
<thead>
  <tr><th colspan="6">Single-nodetunggal- GPU/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Presisi</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>128</td><td>112</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>128</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>128</td><td>135</td></tr>
  <tr><td>albert-base-v2</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>191</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>64</td><td>94</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>96</td><td>101</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>96</td><td>96</td></tr>
  <tr><td>bert-base-uncased</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>128</td></tr>
  <tr><td>bert-large-uncased</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>35</td><td>21</td></tr>
  <tr><td>bert-large-uncased</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>39</td><td>26</td></tr>
  <tr><td>bert-large-uncased</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>60</td><td>50</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>96</td><td>90</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>96</td><td>98</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>96</td><td>96</td></tr>
  <tr><td>camembert-base</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>128</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>256</td><td>160</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>128</td><td>176</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>128</td><td>160</td></tr>
  <tr><td>distilbert-base-uncased</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>256</td><td>258</td></tr>
  <tr><td>google_electra-diskriminator kecil</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>256</td><td>216</td></tr>
  <tr><td>google_electra-diskriminator kecil</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>256</td><td>230</td></tr>
  <tr><td>google_electra-diskriminator kecil</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>256</td><td>224</td></tr>
  <tr><td>google_electra-diskriminator kecil</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>256</td><td>320</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>80</td><td>64</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>80</td><td>77</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>80</td><td>72</td></tr>
  <tr><td>gpt2</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>120</td></tr>
  <tr><td>jplu_tf-xlm-roberta-base</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>28</td><td>24</td></tr>
  <tr><td>jplu_tf-xlm-roberta-base</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>32</td><td>24</td></tr>
  <tr><td>jplu_tf-xlm-roberta-base</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>32</td><td>26</td></tr>
  <tr><td>jplu_tf-xlm-roberta-base</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>66</td><td>52</td></tr>
  <tr><td>microsoft_mpnet-basis</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>96</td><td>92</td></tr>
  <tr><td>microsoft_mpnet-basis</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>96</td><td>101</td></tr>
  <tr><td>microsoft_mpnet-basis</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>96</td><td>101</td></tr>
  <tr><td>microsoft_mpnet-basis</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>152</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>g4dn.16xlarge</td><td>mengapung16</td><td>64</td><td>72</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>p3.2xlarge</td><td>mengapung16</td><td>64</td><td>84</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>p3.8xlarge</td><td>mengapung16</td><td>64</td><td>86</td></tr>
  <tr><td>Roberta base</td><td>wikitext-2-raw-v1</td><td>g5.4xlarge</td><td>mengapung16</td><td>128</td><td>128</td></tr>
</tbody>
</table>


### TensorFlow 2.9.1
<a name="training-compiler-tested-models-tf291"></a>

Diuji menggunakan [ TensorFlow Model Garden ](https://github.com/tensorflow/models) dengan Automatic Mixed Precision (AMP).


<table>
<thead>
  <tr><th colspan="5">Single-nodetunggal- GPU/multi-GPU</th></tr>
  <tr><th>Model</th><th>Set data</th><th>Tipe instans</th><th>Ukuran batch untuk kerangka kerja asli </th><th>Ukuran batch untuk Kompil SageMaker er Pelatihan </th></tr>
</thead>
<tbody>
  <tr><td>ResNet50</td><td>ImageNet</td><td>ml.g4dn.2xbesar</td><td>192</td><td>256*</td></tr>
  <tr><td rowspan="3">ResNet101</td><td rowspan="3">ImageNet</td><td>ml.g4dn.2xbesar</td><td>128</td><td>160</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>224</td><td>256*</td></tr>
  <tr><td>ml.p3.16xlarge</td><td>1536</td><td>1792</td></tr>
  <tr><td rowspan="3">ResNet152</td><td rowspan="3">ImageNet</td><td>ml.g5.2xbesar</td><td>192</td><td>224</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>160</td><td>160</td></tr>
  <tr><td>ml.p3.16xlarge</td><td>1024</td><td>1280</td></tr>
  <tr><td rowspan="4">VisionTransformer</td><td rowspan="4">ImageNet</td><td>ml.g4dn.2xbesar</td><td>80</td><td>128*</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>112</td><td>128*</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>56</td><td>128*</td></tr>
  <tr><td>ml.p3.16xlarge</td><td>640</td><td>1024*</td></tr>
  <tr><td rowspan="4">DetectionTransformer-ResNet50</td><td rowspan="4">COCO-2017</td><td>ml.g4dn.2xbesar</td><td>2</td><td>2</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>3</td><td>6</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>2</td><td>4</td></tr>
  <tr><td>ml.p3.16xlarge</td><td>8</td><td>32</td></tr>
  <tr><td rowspan="3">MaskRCNN-ResNet50-FPN</td><td rowspan="3">COCO-2017</td><td>ml.g4dn.2xbesar</td><td>4</td><td>4</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>6</td><td>8</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>4</td><td>6</td></tr>
</tbody>
</table>


\* Ukuran batch yang ditandai dengan simbol tanda bintang (\*) menunjukkan ukuran batch terbesar yang diuji oleh tim pengembang SageMaker Training Compiler. Untuk sel yang ditandai, instance mungkin dapat menyesuaikan ukuran batch yang lebih besar dari yang ditunjukkan.

### Transformer 4.21.1 dengan 1.11.0 PyTorch
<a name="training-compiler-tested-models-hf421-pt111"></a>

Diuji dengan `Sequence_Len=512` dan Precision Campuran Otomatis (AMP).


<table>
<thead>
  <tr><th colspan="6">Single-node GPU tunggal</th></tr>
  <tr><th>Model </th><th>Set data</th><th>Tipe instans</th><th>Jumlah instans</th><th>Ukuran batch untuk kerangka kerja asli</th><th>Ukuran batch untuk Kompiler Pelatihan</th></tr>
</thead>
<tbody>
  <tr><td rowspan="3">albert-base-v2</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>14</td><td>28</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>18</td><td>40</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>14</td><td>32</td></tr>
  <tr><td rowspan="3">bert-base-casing</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>12</td><td>24</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>28</td><td>44</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>16</td><td>20</td></tr>
  <tr><td rowspan="3">camembert-base</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>16</td><td>28</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>24</td><td>40</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>16</td><td>24</td></tr>
  <tr><td rowspan="4">distilbert-base-uncased</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>28</td><td>52</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>40</td><td>76</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>32</td><td>48</td></tr>
  <tr><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>4</td><td>82</td><td>160</td></tr>
  <tr><td rowspan="3">distilgpt2</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>6</td><td>18</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>12</td><td>28</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>6</td><td>16</td></tr>
  <tr><td rowspan="3">distilroberta-base</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>20</td><td>40</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>28</td><td>56</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>24</td><td>40</td></tr>
  <tr><td rowspan="3">EleutherAI/gpt-neo-125M</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>4</td><td>8</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>6</td><td>14</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>4</td><td>10</td></tr>
  <tr><td rowspan="4">gpt2</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>4</td><td>8</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>6</td><td>16</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>4</td><td>10</td></tr>
  <tr><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>4</td><td>13</td><td>25</td></tr>
  <tr><td rowspan="4">Roberta base</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>12</td><td>20</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>24</td><td>36</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>12</td><td>20</td></tr>
  <tr><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>4</td><td>36</td><td>64</td></tr>
  <tr><td rowspan="3">xlnet-base casing</td><td rowspan="3">wikiteks-2</td><td>ml.g4dn.2xbesar</td><td>1</td><td>2</td><td>6</td></tr>
  <tr><td>ml.g5.2xbesar</td><td>1</td><td>2</td><td>10</td></tr>
  <tr><td>ml.p3.2xlarge</td><td>1</td><td>2</td><td>8</td></tr>
  <tr><td rowspan="4">bert-base-uncased</td><td rowspan="4">wikitext-103-v1</td><td rowspan="4">ml.p4d.24xbesar</td><td>2</td><td>32</td><td>64</td></tr>
  <tr><td>4</td><td>32</td><td>64</td></tr>
  <tr><td>8</td><td>32</td><td>64</td></tr>
  <tr><td>16</td><td>32</td><td>64</td></tr>
  <tr><td>roberta-besar</td><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>4</td><td>16</td><td>24</td></tr>
  <tr><td>microsoft/deberta-v3-basis</td><td>wikitext-103-v1</td><td>ml.p4d.24xbesar</td><td>16</td><td>9</td><td>23</td></tr>
</tbody>
</table>


### Transformers 4.17.0 dengan 1.10.2 PyTorch
<a name="training-compiler-tested-models-hf417-pt110"></a>

Diuji dengan `Sequence_Len=512` dan Precision Campuran Otomatis (AMP).


<table>
<thead>
  <tr><th colspan="4">Single-node GPU tunggal</th></tr>
  <tr><th>Model </th><th>Tipe instans</th><th>Ukuran batch untuk kerangka kerja asli</th><th>Ukuran batch untuk Kompiler Pelatihan</th></tr>
</thead>
<tbody>
  <tr><td rowspan="2">albert-base-v2</td><td>ml.p3.2xlarge</td><td>14</td><td>28</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>14</td><td>24</td></tr>
  <tr><td rowspan="2">bert-base-casing</td><td>ml.p3.2xlarge</td><td>16</td><td>24</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>12</td><td>24</td></tr>
  <tr><td rowspan="2">bert-base-uncased</td><td>ml.p3.2xlarge</td><td>16</td><td>24</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>12</td><td>28</td></tr>
  <tr><td rowspan="2">camembert-base</td><td>ml.p3.2xlarge</td><td>12</td><td>24</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>12</td><td>28</td></tr>
  <tr><td rowspan="2">distilbert-base-uncased</td><td>ml.p3.2xlarge</td><td>28</td><td>48</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>24</td><td>52</td></tr>
  <tr><td rowspan="2">distilgpt2</td><td>ml.p3.2xlarge</td><td>6</td><td>12</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>6</td><td>14</td></tr>
  <tr><td rowspan="2">distilroberta-base</td><td>ml.p3.2xlarge</td><td>20</td><td>40</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>12</td><td>40</td></tr>
  <tr><td rowspan="2">EleutherAI/gpt-neo-125M</td><td>ml.p3.2xlarge</td><td>2</td><td>10</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>2</td><td>8</td></tr>
  <tr><td rowspan="2">facebook/bart-basis</td><td>ml.p3.2xlarge</td><td>2</td><td>6</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>2</td><td>6</td></tr>
  <tr><td rowspan="2">gpt2</td><td>ml.p3.2xlarge</td><td>4</td><td>8</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>2</td><td>8</td></tr>
  <tr><td rowspan="2">Roberta base</td><td>ml.p3.2xlarge</td><td>12</td><td>20</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>12</td><td>20</td></tr>
  <tr><td rowspan="2">xlnet-base casing</td><td>ml.p3.2xlarge</td><td>2</td><td>8</td></tr>
  <tr><td>ml.g4dn.2xbesar</td><td>4</td><td>6</td></tr>
</tbody>
</table>


### Transformers 4.11.0 dengan 1.9.0 PyTorch
<a name="training-compiler-tested-models-hf411-pt190"></a>

Diuji dengan `Sequence_Len=512` dan Precision Campuran Otomatis (AMP).


<table>
<thead>
  <tr><th colspan="4">Single-node GPU tunggal</th></tr>
  <tr><th>Model </th><th>Tipe instans</th><th>Ukuran batch untuk asli</th><th>Ukuran batch untuk Kompiler Pelatihan</th></tr>
</thead>
<tbody>
  <tr><td>albert-base-v2 </td><td>ml.p3.2xlarge</td><td>12</td><td>32</td></tr>
  <tr><td>bert-base-casing </td><td>ml.p3.2xlarge</td><td>14</td><td>24</td></tr>
  <tr><td>bert-base-china</td><td>ml.p3.2xlarge</td><td>16</td><td>24</td></tr>
  <tr><td>bert-base-berkasing multibahasa </td><td>ml.p3.2xlarge</td><td>4</td><td>16</td></tr>
  <tr><td>bert-base-multilingual-uncased </td><td>ml.p3.2xlarge</td><td>8</td><td>16</td></tr>
  <tr><td>bert-base-uncased </td><td>ml.p3.2xlarge</td><td>12</td><td>24</td></tr>
  <tr><td>cl- tohoku/bert -basis-jepangan-penyamaran kata utuh</td><td>ml.p3.2xlarge</td><td>12</td><td>24</td></tr>
  <tr><td>cl- tohoku/bert -base-jepang </td><td>ml.p3.2xlarge</td><td>12</td><td>24</td></tr>
  <tr><td>distilbert-base-uncased </td><td>ml.p3.2xlarge</td><td>28</td><td>32</td></tr>
  <tr><td>distiler-base-uncased-finetuned-sst-2-english</td><td>ml.p3.2xlarge</td><td>28</td><td>32</td></tr>
  <tr><td>distilgpt2 </td><td>ml.p3.2xlarge</td><td>16</td><td>32</td></tr>
  <tr><td>facebook/bart-basis </td><td>ml.p3.2xlarge</td><td>4</td><td>8</td></tr>
  <tr><td>gpt2</td><td>ml.p3.2xlarge</td><td>6</td><td>20</td></tr>
  <tr><td>nreimers/MiniLMv2-L6-H384-distilled-from-RoBERTa-Large </td><td>ml.p3.2xlarge</td><td>20</td><td>32</td></tr>
  <tr><td>Roberta base </td><td>ml.p3.2xlarge</td><td>12</td><td>20</td></tr>
</tbody>
</table>



<table>
<thead>
  <tr><th colspan="4">Single-node Multi-GPU</th></tr>
  <tr><th>Model </th><th>Tipe instans</th><th>Ukuran batch untuk asli</th><th>Ukuran batch untuk Kompiler Pelatihan</th></tr>
</thead>
<tbody>
  <tr><td>bert-base-china </td><td>ml.p3.8xlarge</td><td>16</td><td>26</td></tr>
  <tr><td>bert-base-berkasing multibahasa </td><td>ml.p3.8xlarge</td><td>6</td><td>16</td></tr>
  <tr><td>bert-base-multilingual-uncased</td><td>ml.p3.8xlarge</td><td>6</td><td>16</td></tr>
  <tr><td>bert-base-uncased </td><td>ml.p3.8xlarge</td><td>14</td><td>24</td></tr>
  <tr><td>distilbert-base-uncased </td><td>ml.p3.8xlarge</td><td>14</td><td>32</td></tr>
  <tr><td>distilgpt2</td><td>ml.p3.8xlarge</td><td>6</td><td>32</td></tr>
  <tr><td>facebook/bart-basis</td><td>ml.p3.8xlarge</td><td>8</td><td>16</td></tr>
  <tr><td>gpt2 </td><td>ml.p3.8xlarge</td><td>8</td><td>20</td></tr>
  <tr><td>Roberta base </td><td>ml.p3.8xlarge</td><td>12</td><td>20</td></tr>
</tbody>
</table>


### Transformers 4.17.0 dengan 2.6.3 TensorFlow
<a name="training-compiler-tested-models-hf417-tf263"></a>

Diuji dengan `Sequence_Len=128` dan Precision Campuran Otomatis (AMP).


| Model  | Tipe instans | Ukuran batch untuk kerangka kerja asli | Ukuran batch untuk Kompiler Pelatihan | 
| --- | --- | --- | --- | 
| albert-base-v2 | ml.g4dn.16xbesar | 136 | 208 | 
| albert-base-v2 | ml.g5.4xbesar | 219 | 312 | 
| albert-base-v2 | ml.p3.2xlarge | 152 | 208 | 
| albert-base-v2 | ml.p3.8xlarge | 152 | 192 | 
| bert-base-uncased | ml.g4dn.16xbesar | 120 | 101 | 
| bert-base-uncased | ml.g5.4xbesar | 184 | 160 | 
| bert-base-uncased | ml.p3.2xlarge | 128 | 108 | 
| bert-large-uncased | ml.g4dn.16xbesar | 37 | 28 | 
| bert-large-uncased | ml.g5.4xbesar | 64 | 55 | 
| bert-large-uncased | ml.p3.2xlarge | 40 | 32 | 
| camembert-base | ml.g4dn.16xbesar | 96 | 100 | 
| camembert-base | ml.g5.4xbesar | 190 | 160 | 
| camembert-base | ml.p3.2xlarge | 129 | 108 | 
| camembert-base | ml.p3.8xlarge | 128 | 104 | 
| distilbert-base-uncased | ml.g4dn.16xbesar | 210 | 160 | 
| distilbert-base-uncased | ml.g5.4xbesar | 327 | 288 | 
| distilbert-base-uncased | ml.p3.2xlarge | 224 | 196 | 
| distilbert-base-uncased | ml.p3.8xlarge | 192 | 182 | 
| google\_electra-diskriminator kecil | ml.g4dn.16xbesar | 336 | 288 | 
| google\_electra-diskriminator kecil | ml.g5.4xbesar | 504 | 384 | 
| google\_electra-diskriminator kecil | ml.p3.2xlarge | 352 | 323 | 
| gpt2 | ml.g4dn.16xbesar | 89 | 64 | 
| gpt2 | ml.g5.4xbesar | 140 | 146 | 
| gpt2 | ml.p3.2xlarge | 94 | 96 | 
| gpt2 | ml.p3.8xlarge | 96 | 88 | 
| jplu\_tf-xlm-roberta-base | ml.g4dn.16xbesar | 52 | 16 | 
| jplu\_tf-xlm-roberta-base | ml.g5.4xbesar | 64 | 44 | 
| microsoft\_mpnet-basis | ml.g4dn.16xbesar | 120 | 100 | 
| microsoft\_mpnet-basis | ml.g5.4xbesar | 192 | 160 | 
| microsoft\_mpnet-basis | ml.p3.2xlarge | 128 | 104 | 
| microsoft\_mpnet-basis | ml.p3.8xlarge | 130 | 92 | 
| Roberta base | ml.g4dn.16xbesar | 108 | 64 | 
| Roberta base | ml.g5.4xbesar | 176 | 142 | 
| Roberta base | ml.p3.2xlarge | 118 | 100 | 
| Roberta base | ml.p3.8xlarge | 112 | 88 | 

### Transformers 4.11.0 dengan 2.5.1 TensorFlow
<a name="training-compiler-tested-models-hf411-tf251"></a>

Diuji dengan `Sequence_Len=128` dan Precision Campuran Otomatis (AMP).


<table>
<thead>
  <tr><th colspan="4">Single-node GPU tunggal</th></tr>
  <tr><th>Model </th><th>Tipe instans</th><th>Ukuran batch untuk asli</th><th>Ukuran batch untuk Kompiler Pelatihan</th></tr>
</thead>
<tbody>
  <tr><td>albert-base-v2 </td><td>ml.p3.2xlarge</td><td>128</td><td>128</td></tr>
  <tr><td>bart-base </td><td>ml.p3.2xlarge</td><td>12</td><td>64</td></tr>
  <tr><td>bart-besar </td><td>ml.p3.2xlarge</td><td>4</td><td>28</td></tr>
  <tr><td>bert-base-casing </td><td>ml.p3.2xlarge</td><td>16</td><td>128</td></tr>
  <tr><td>bert-base-china</td><td>ml.p3.2xlarge</td><td>16</td><td>128</td></tr>
  <tr><td>bert-base-berkasing multibahasa </td><td>ml.p3.2xlarge</td><td>12</td><td>64</td></tr>
  <tr><td>bert-base-multilingual-uncased </td><td>ml.p3.2xlarge</td><td>16</td><td>96</td></tr>
  <tr><td>bert-base-uncased</td><td>ml.p3.2xlarge</td><td>16</td><td>96</td></tr>
  <tr><td>bert-large-uncased </td><td>ml.p3.2xlarge</td><td>4</td><td>24</td></tr>
  <tr><td>cl- tohoku/bert -base-jepang </td><td>ml.p3.2xlarge</td><td>16</td><td>128</td></tr>
  <tr><td>cl- tohoku/bert -basis-jepangan-penyamaran kata utuh </td><td>ml.p3.2xlarge</td><td>16</td><td>128</td></tr>
  <tr><td>distiler-base-sst2 </td><td>ml.p3.2xlarge</td><td>32</td><td>128</td></tr>
  <tr><td>distilbert-base-uncased </td><td>ml.p3.2xlarge</td><td>32</td><td>128</td></tr>
  <tr><td>distilgpt2</td><td>ml.p3.2xlarge</td><td>32</td><td>128</td></tr>
  <tr><td>gpt2 </td><td>ml.p3.2xlarge</td><td>12</td><td>64</td></tr>
  <tr><td>gpt2-besar </td><td>ml.p3.2xlarge</td><td>2</td><td>24</td></tr>
  <tr><td>jplu/tf-xlm-roberta-basis </td><td>ml.p3.2xlarge</td><td>12</td><td>32</td></tr>
  <tr><td>Roberta base </td><td>ml.p3.2xlarge</td><td>4</td><td>64</td></tr>
  <tr><td>roberta-besar </td><td>ml.p3.2xlarge</td><td>4</td><td>64</td></tr>
  <tr><td>t5-basis </td><td>ml.p3.2xlarge</td><td>64</td><td>64</td></tr>
  <tr><td>t5-kecil </td><td>ml.p3.2xlarge</td><td>128</td><td>128</td></tr>
</tbody>
</table>