Ecosyste.ms: Timeline
Browse the timeline of events for every public repo on GitHub. Data updated hourly from GH Archive.
github-actions[bot] pushed 1 commit to gh-pages ai4co/rl4co
- Deployed 28c1166 with MkDocs version: 1.6.1 e8cf947
fedebotu created a comment on an issue on ai4co/rl4co
I tried manually updating all dependencies, and it turns out that, on a 3090, the bug may be reproducible. In my case, the loss explodes during the Epoch 2 on [this notebook](https://github.com/ai4...
github-actions[bot] pushed 1 commit to gh-pages ai4co/rl4co
- Deployed 3f9eb35 with MkDocs version: 1.6.1 e7dfb69
fedebotu pushed 8 commits to main ai4co/rl4co
- [Minor] update ignore CO-related filetypes 8a4920e
- [Minor] update ignore CO-related filetypes 697f45e
- [Chore] update spec names 291a65a
- [Version] 0.5.1; update dependencies cf52447
- [Chore] update spec names 74b18bd
- [Example] CVRPLib update 7a16b64
- [Example] TSPLib update acb4b3d
- [Version] 0.5.1dev0 3f9eb35
ShuN6211 closed an issue on ai4co/rl4co
[BUG] Tensordict v0.6.0 is not compatible with rl4co
## Describe the bug Tensordictv0.6,0 was released in last night, but this is not compatible with rl4co. https://github.com/pytorch/tensordict/releases/tag/v0.6.0 ## To Reproduce ```bash cd...ShuN6211 created a comment on an issue on ai4co/rl4co
@fedebotu @cbhua It has been a few days since TorchRL v0.6.0 was released, there is no troublesome in my end. If you are also have any problem, I would like to say let's close this issue.
alopezrivera created a comment on an issue on ai4co/rl4co
So far I was able to reproduce this behavior on the 4 systems below (all on RunPod), which makes me think the environment is the culprit here - L40, `pytorch:2.4.0-py3.11-cuda12.4.1-devel-ubuntu22...
fedebotu created a comment on an issue on ai4co/rl4co
Good, also you may try testing it in another system. For example if the problem is reproducible in Google Colab, then it is likely an important issue in RL4CO. Otherwise it might be outside of our ...
alopezrivera created a comment on an issue on ai4co/rl4co
In that case it is highly likely that this had to do with either numerical precision settings or some operation conducted inside my space vehicle routing environment. I'll look into those two possi...
github-actions[bot] pushed 1 commit to gh-pages ai4co/rl4co
- Deployed adac2f6 with MkDocs version: 1.6.1 5d88a5f