metadata
language: en
tags:
- sagemaker
- bart
- summarization
license: apache-2.0
datasets:
- samsum
widget:
- text: >-
Jeff: Can I train a 🤗 Transformers model on Amazon SageMaker?
Philipp: Sure you can use the new Hugging Face Deep Learning Container.
Jeff: ok.
Jeff: and how can I get started?
Jeff: where can I find documentation?
Philipp: ok, ok you can find everything here.
https://huggingface.co./blog/the-partnership-amazon-sagemaker-and-hugging-face
model-index:
- name: philschmid/distilbart-cnn-12-6-samsum
results:
- task:
type: summarization
name: Summarization
dataset:
name: samsum
type: samsum
config: samsum
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 41.0895
verified: true
- name: ROUGE-2
type: rouge
value: 20.7459
verified: true
- name: ROUGE-L
type: rouge
value: 31.5952
verified: true
- name: ROUGE-LSUM
type: rouge
value: 38.3389
verified: true
- name: loss
type: loss
value: 1.4566329717636108
verified: true
- name: gen_len
type: gen_len
value: 59.6032
verified: true
- task:
type: summarization
name: Summarization
dataset:
name: xsum
type: xsum
config: default
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 21.1644
verified: true
- name: ROUGE-2
type: rouge
value: 4.0659
verified: true
- name: ROUGE-L
type: rouge
value: 13.9414
verified: true
- name: ROUGE-LSUM
type: rouge
value: 17.0718
verified: true
- name: loss
type: loss
value: 3.002755880355835
verified: true
- name: gen_len
type: gen_len
value: 71.2969
verified: true
distilbart-cnn-12-6-samsum
This model was trained using Amazon SageMaker and the new Hugging Face Deep Learning container.
For more information look at:
- 🤗 Transformers Documentation: Amazon SageMaker
- Example Notebooks
- Amazon SageMaker documentation for Hugging Face
- Python SDK SageMaker documentation for Hugging Face
- Deep Learning Container
Hyperparameters
{
"dataset_name": "samsum",
"do_eval": true,
"do_train": true,
"fp16": true,
"learning_rate": 5e-05,
"model_name_or_path": "sshleifer/distilbart-cnn-12-6",
"num_train_epochs": 3,
"output_dir": "/opt/ml/model",
"per_device_eval_batch_size": 8,
"per_device_train_batch_size": 8,
"seed": 7
}
Train results
| key | value |
|---|---|
| epoch | 3.0 |
| init_mem_cpu_alloc_delta | 180338 |
| init_mem_cpu_peaked_delta | 18282 |
| init_mem_gpu_alloc_delta | 1222242816 |
| init_mem_gpu_peaked_delta | 0 |
| train_mem_cpu_alloc_delta | 6971403 |
| train_mem_cpu_peaked_delta | 640733 |
| train_mem_gpu_alloc_delta | 4910897664 |
| train_mem_gpu_peaked_delta | 23331969536 |
| train_runtime | 155.2034 |
| train_samples | 14732 |
| train_samples_per_second | 2.242 |
Eval results
| key | value |
|---|---|
| epoch | 3.0 |
| eval_loss | 1.4209576845169067 |
| eval_mem_cpu_alloc_delta | 868003 |
| eval_mem_cpu_peaked_delta | 18250 |
| eval_mem_gpu_alloc_delta | 0 |
| eval_mem_gpu_peaked_delta | 328244736 |
| eval_runtime | 0.6088 |
| eval_samples | 818 |
| eval_samples_per_second | 1343.647 |
Usage
from transformers import pipeline
summarizer = pipeline("summarization", model="philschmid/distilbart-cnn-12-6-samsum")
conversation = '''Jeff: Can I train a 🤗 Transformers model on Amazon SageMaker?
Philipp: Sure you can use the new Hugging Face Deep Learning Container.
Jeff: ok.
Jeff: and how can I get started?
Jeff: where can I find documentation?
Philipp: ok, ok you can find everything here. https://huggingface.co./blog/the-partnership-amazon-sagemaker-and-hugging-face
'''
nlp(conversation)