Spaces:
Running
on
CPU Upgrade
Unable to submit model
The model and tokenizer have been loaded via AutoClasses.
However, the following error occurs.
Model "4season/alignment-model-test2" 's tokenizer cannot be loaded. Is your tokenizer class in a stable transformers release, and correctly configured?
used LlamaTokenizer.
Also, "4season/alignment_model_test" was not originally classified as a merge model, but I do not know why it was suddenly classified as a merge model.
Hi, thanks for your issue!
For 4season/alignment_model_test, you say in your model card that After DPO training, we linearly merged models to boost performance.. So your model is a merge and was classified as such.
We'll investigate for the tokenizer of 4season/alignment-model-test2, but you've changed your Tokenizer configuration quite a lot - with which version of your model did you try?
Tried the most Recent version.
Hi!
Your tokenizer has a problem and is not configured properly. (Please follow the Submission steps next time, and make sure your model and tokenizer are loadable with AutoClasses.)
If you try to do
tk = AutoTokenizer.from_pretrained(
"4season/alignment-model-test2",
revision="main",
trust_remote_code=False,
token=None
)
it fails, with the error TypeError: not a string.
Closing this issue - feel free to try to resubmit once you fixed your tokenizer!