dolly: HFValidationError: Repo id must be in the form 'repo_name' or 'namespace/repo_name'
deepspeed training error, how to deal with it.
!deepspeed --num_gpus=1
–module training.trainer
–input-model /content/drive/MyDrive/EleutherAI/pythia-6.9b
–deepspeed config/ds_z3_bf16_config.json
–epochs 2
–local-output-dir /content/drive/MyDrive/dolly/training_out
–dbfs-output-dir /content/drive/MyDrive/dolly/dolly_training
–per-device-train-batch-size 6
–per-device-eval-batch-size 6
–logging-steps 10
–save-steps 200
–save-total-limit 20
–eval-steps 50
–warmup-steps 50
–test-size 200
–lr 5e-6
HFValidationError: Repo id must be in the form ‘repo_name’ or ‘namespace/repo_name’: ‘/content/drive/MyDrive/EleutherAI/pythia-6.9b’. Use repo_type argument if needed.
About this issue
- Original URL
- State: closed
- Created a year ago
- Comments: 23 (1 by maintainers)
Are you sure that path exists and contains the model? Usually I’ve seen that when the model does not exist at that path or some expected files are missing.
If you’re trying to load the model checkpoints from local machine, give the FULL directory path where the model checkpoints are saved. FYI, passing the relative path like: “
~/models/7B” => fails. Giving the full path like this: “/home/users/abc/models/7B” => this works fine. This worked for me.Just pass
--input-model EleutherAI/pythia-6.9bCorrect, you use /dbfs not dbfs: with tools that only know local files. But I am not clear what you are saying the root cause was