File size: 941 Bytes
586080e 03deb4c 586080e 03deb4c f7b94b6 |
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 |
---
license: mit
language:
- de
tags:
- title generation
- headline-generation
- teaser generation
- keyword generation
- tweet generation
- news
inference: false
---
# snip-igel-100
<!-- Provide a quick summary of what the model is/does. -->
snip-igel-100
Version 1.0 / 13 April 2023
An adapter for [IGEL](https://huggingface.co/philschmid/instruct-igel-001) to generate german news snippets with human written instructions
See [snip-igel-500](https://huggingface.co/snipaid/snip-igel-500) for the full model description. We repeated fine-tuning with gradually increased amounts of training data, to see the difference.
# Environmental Impact
Carbon emissions were estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact/#compute) presented in Lacoste et al. (2019).
Hardware Type: RTX 4090
Hours used: 21min 57s
Cloud Provider: Vast.ai
Compute Region: Poland
Carbon Emitted: ~0.11 kg of CO2e |