evolu8

garder14/byol-tensorflow2 (batch-norm & softmax/cross-entropy)

1

Running TF 2.4.1 with seeds and envs set I'm getting different results each run for this guy: https://github.com/garder14/byol-tensorflow2 I currently suspect it's the gradient tape. Not sure how to handle...

Host the server myself?

2

Hi This would be wonderful if I could host the service on my own hardware. Is the server something you intend to open source?

Training fails

3

Running training results in the following: ``` Error executing job with overrides: [] Traceback (most recent call last): File "train_segmentation.py", line 532, in my_app pos_labels=True File "/home/ec2-user/SageMaker/STEGO/src/core.py", line 740, in...

Only the first checkpoint link works

1

Are the links for pretrained models beyond the first intentionally dead? Would love access to the Large Q&A version. Is there an ETA if 'coming soon'? Best

Use local model

38

### 🚀 The feature It would be wonderful to use a locally hosted StarCoder or MPT model. As opposed you require an HF or OpenAI key. ### Motivation, pitch Data...

I'm using gcloud to run this guy locally (all I really wont is to be able to use scipy), it creates the vm, the health checker 200s but any call...

reset weights

1

Would be nice to be able to restart training with reinitialised weights, quickly and easily. Not sure how best to do this right now.

GPU requirements (and other dependancies)

32

Hi It would be great to include GPU RAM and storage requirements in the README. The requirements appear to be non-trivial for many. Great work. Thank you for opening! It...

Llama2

2

Loving med-flamingo. However, would also love to use Llama2. Is this likely to come?

### Check Existing Issues - [x] I have searched the existing issues and discussions. ### Problem Description Jupyter runs with a default, global kernel. This is often not safe or...