Hello, what version of COMET should I use to use Unbabel/wmt22-cometkiwi-da? I used the Hugging Face posted code and got KeyError: "Model 'Unbabel/wmt22-cometkiwi-da' not supported by COMET."
Hello, what version of COMET should I use to use Unbabel/wmt22-cometkiwi-da? I used the Hugging Face posted code and got KeyError: "Model 'Unbabel/wmt22-cometkiwi-da' not supported by COMET."
CometKiwi models should run on unbabel-comet>=2.0. Your error is most probably because you have not acknowledge it's license. You have to acknowledge its license and then run the following:
huggingface-cli login
# or using an environment variable
huggingface-cli login --token $HUGGINGFACE_TOKEN
Then you should be able to run the model.
Hi,
I have a similar problem. My comet version is "version = "2.2.2" " and I use Python 3.11.
I have accepted the license and logged in to Huggingface on my computer.
Still, I receive the error message..
My comet version is "version = "2.2.2" " and I use Python 3.10.12
pip install --upgrade pip pip install unbabel-comet from comet import download_model, load_from_checkpoint model_path = download_model("Unbabel/wmt22-comet-da") model = load_from_checkpoint(model_path) data = [ { "src": "Dem Feuer konnte Einhalt geboten werden", "mt": "The fire could be stopped", "ref": "They were able to control the fire." }, { "src": "Schulen und Kindergärten wurden eröffnet.", "mt": "Schools and kindergartens were open", "ref": "Schools and kindergartens opened" } ] model_output = model.predict(data, batch_size=8, gpus=1) print (model_output)
I checked my code and it still works.there is the demo ,maybe you can have a try.
Hi,
I realized that my token was missing the permission to read gated models:It now works :)
This is strange. I am using Comet version 2.2.2 and Python 3.11, and I have also checked this option, but I still encounter the above error.
I realized where I went wrong. Previously, when loading the model, the path was set to the folder containing the model. Changing the model path to path/wmt22-cometkiwi-da/checkpoints/model.ckpt resolved the issue.