thomasmdrayton
thomasmdrayton
Apologies not sure why the previous code block did not render: train.py: ``` import argparse import os import pandas as pd import numpy as np import torch import wandb from...
Hi @Sascha-Roe, I'm just following up to see if the solution worked for you?
Hi @Daraan, thanks for the detailed reproduction! It's much appreciated. Would you also mind providing the `wandb` sdk version you are using for completeness!
Thanks for the extra detail @Daraan - as you're aware, there is currently a bug with forking runs so I'll pick this up when it is first resolved - thanks...
Hi @Daraan, just to mention, I'll update you once I hear back from our engineers on the bug!
Hi @Daraan, thanks for reaching out with the feature request and your feedback. It is really appreciated! Once we've resolved the current issue we're aware of with forking, I'll pick...
Hi @Daraan , can you clarify if you are using a GraphQL query to access the branch point?
Thanks for the clarity! I've created a feature request. It `branch_point` dissappears but the run_id remains the same for a resumed run. I guess the workaround for the time being...
Hi @Daraan, just to check in here - and apologies for the delay! Although the `run_id` doesn't change when you resume a run, would you still like to be able...
Thanks for confirming! We'll keep you posted!