How to load custom-trained model and inference?

Thank you for releasing the training code. I have train a model and I am wondering how to load it on the inference stage.
Could you please give me some example scripts or advice?

The checkpoints strcture:
exps/..../000100/
custom_checkpoint_0.pkl model.safetensors optimizer.bin random_states_0.pkl

Hi there,
Thanks for your interest!

Before running the inference script, you should first convert the training checkpoint to huggingface-compatible models by running python scripts/convert_hf.py --config <YOUR_EXACT_TRAINING_CONFIG>. The converted checkpoint will be saved under exps/releases.

Plz feel free to comment if there is still any trouble doing inference.

Proxy causes error. Wondering how to solve...

Traceback (most recent call last):
  File "OpenLRM/scripts/convert_hf.py", line 80, in <module>
    loaded_step = auto_load_model(cfg, hf_model)
  File "OpenLRM/openlrm/utils/proxy.py", line 38, in wrapper
    os.environ['HTTP_PROXY'] = HTTP_PROXY
  File "/usr/local/lib/python3.10/os.py", line 685, in __setitem__
    value = self.encodevalue(value)
  File "/usr/local/lib/python3.10/os.py", line 757, in encode
    raise TypeError("str expected, not %s" % type(value).__name__)
TypeError: str expected, not NoneType

Hi,

Plz comment out this line as a workaround here.

I'll try fix this problem later by detecting special environ vars and avoiding calling no_proxy by default.

Hi,

Fixed in this commit here.

OpenLRM/openlrm/utils/proxy.py

Line 18 in c2260e0

NO_PROXY = "OPENLRM_NO_DATA_PROXY" in os.environ

It should work now without manually commenting out no_proxy.

Thank you for releasing the training code. I have train a model and I am wondering how to load it on the inference stage. Could you please give me some example scripts or advice?

The checkpoints strcture: exps/..../000100/ custom_checkpoint_0.pkl model.safetensors optimizer.bin random_states_0.pkl

where did you configure the dataset path? what was the folder structure of the dataset?