Arthur

Hey! I'll try to have a look, it looks like setting the `vocab_size` does not change the shape of the model which means that it does not raise an error...

Hey! There must be a typo in your `generation_config` as the `convert_llama_weights_to_hf.py` as well as `configuration_llama` both set it to `2`. Are you sure that you are using the latest...

Indeed, just realised you have to `cat` them 😞 my bad!

Hey! Thanks for reporting. Could you share the audio that you are using? We never stumbled upon something like this 😅 Problem seems to come from `sequence[1:relevant_timestamp]` ,but the traceback...

Hey! Thanks for reporting. This is normal as the `pipeline` does not support returning the usual `dictionary`. We should probably prevent this behaviour (raise an error when `return_dict_in_generate` is set...

Best recommendation in the mean time is to define a custom pipeline, where you process the inputs before feeding them to `super.preprocess`!

Hey thanks for reporting!

Hey! Thanks for reporting. This is indeed a bug, will see what we can do to fix that!

Hey! Did you try playing with `generation_config` ? All the arguments that you are looking for can either be setup inside, or provided in the `generate` kwargs. Tempertature and penalty...

Hey! As you mention, you are creating an empty directory with the same name. The priority is given to local folders, which explains why you are having this issue. Either...