|
--- |
|
title: README |
|
emoji: 👁 |
|
colorFrom: pink |
|
colorTo: indigo |
|
sdk: static |
|
pinned: false |
|
--- |
|
|
|
<!-- The classes below are necessary for correct rendering --> |
|
<div class="lg:col-span-3"> |
|
<p class="mb-2"> |
|
This organization is a part of the NeurIPS 2021 demonstration <u><a href="https://training-transformers-together.github.io/">"Training Transformers Together"</a></u>. |
|
</p> |
|
<p class="mb-2"> |
|
In this demo, we've trained a model similar to <u><a target="_blank" href="https://openai.com/blog/dall-e/">OpenAI DALL-E</a></u> — |
|
a Transformer "language model" that generates images from text descriptions. |
|
Training happened collaboratively — volunteers from all over the Internet contributed to the training using hardware available to them. |
|
We used <u><a target="_blank" href="https://laion.ai/laion-400-open-dataset/">LAION-400M</a></u>, |
|
the world's largest openly available image-text-pair dataset with 400 million samples. Our model was based on |
|
the <u><a target="_blank" href="https://github.com/lucidrains/DALLE-pytorch">dalle‑pytorch</a></u> implementation |
|
by <u><a target="_blank" href="https://github.com/lucidrains">Phil Wang</a></u> with a few tweaks to make it communication-efficient. |
|
</p> |
|
<p class="mb-2"> |
|
See details about how it works on <u><a target="_blank" href="https://training-transformers-together.github.io/">our website</a></u>. |
|
</p> |
|
<p class="mb-2"> |
|
This organization gathers people participating in the collaborative training and provides links to the related materials: |
|
</p> |
|
<ul class="mb-2"> |
|
<li>👉 <u><a target="_blank" href="https://training-transformers-together.github.io/InferenceResults.html">Inference results</a></u></li></li> |
|
<li>👉 <u><a target="_blank" href="https://huggingface.co/training-transformers-together/dalle-demo-v1">Model weights</a></u> (the latest checkpoint)</li></li> |
|
<li>👉 <u><a target="_blank" href="https://colab.research.google.com/drive/1sXHqy5hKVEQyFX-H2Ai7KzLij-7M_xCB?usp=sharing">Colab notebook for running inference</a> (updated on Apr 5)</u> |
|
<li>👉 <u><a target="_blank" href="https://github.com/learning-at-home/dalle-hivemind">Code</a></u></li> |
|
<li>👉 <u><a target="_blank" href="https://huggingface.co/datasets/laion/laion_100m_vqgan_f8">Dataset</a></u></li> |
|
</ul> |
|
<p class="mb-2"> |
|
The materials below were available during the training run itself: |
|
</p> |
|
<ul class="mb-2"> |
|
<li>👉 Starter kits for <u><a target="_blank" href="https://colab.research.google.com/drive/1BqTWcfsvNQwQqqCRKMKp1_jvQ5L1BhCY?usp=sharing">Google Colab</a></u> and <u><a target="_blank" href="https://www.kaggle.com/yhn112/training-transformers-together/">Kaggle</a></u> (easy way to join the training)</li> |
|
<li>👉 <u><a target="_blank" href="https://huggingface.co/spaces/training-transformers-together/Dashboard">Dashboard</a></u> (the current training state: loss, number of peers, etc.)</li> |
|
<li>👉 Weights & Biases plots for <u><a target="_blank" href="https://wandb.ai/learning-at-home/dalle-hivemind/runs/3l7q56ht">aux peers</a></u> (aggregating the metrics) and actual <u><a target="_blank" href="https://wandb.ai/learning-at-home/dalle-hivemind-trainers">trainers</a></u> (contributing with their GPUs)</li> |
|
</ul> |
|
<p class="mb-2"> |
|
Feel free to reach us on <u><a target="_blank" href="https://discord.gg/uGugx9zYvN">Discord</a></u> if you have any questions 🙂 |
|
</p> |
|
</div> |
|
|
