The HF falcon tutorial has the following line: tokenizer.pad_token = tokenizer.eos_token it looks strange to me. It make sense pad and eos are the same but then why even make a difference between them in the first place in general? cross: Discord hf Why does the falcon QLoRA tutorial code use e…

Personally this is the solution I recommend. But I’ve not tested if Falcon stops producing too many EOSs: def collate_fn_train_only_first_eos_token_mask_everything_after_it(data: list[dict[str, str]], tokenizer: PreTrainedTokenize…

Why does the falcon QLoRA tutorial code use eos_token as pad_token?

Models

brando July 8, 2023, 12:56am 4

Why is this the case? seem really bizzare to me.

Topic		Replies	Views
Mistral trouble when fine-tuning : Don't set pad_token_id = eos_token_id 🤗Transformers	8	5946	August 28, 2024
GPT2 finetuned with eos token will never yield eos token during generation Beginners	6	3384	April 12, 2024
How does GPT decide to stop generating sentences without EOS token? 🤗Transformers	13	24684	August 19, 2024
I want fine tune my LLM (falcon-7b) to learn to stop : Which strategy? Beginners	0	1203	August 9, 2023
Why does hugging face falcon model use mode.config.use_cache = False, why wouldn't it want to have the decoder re-use computations for fine-tuning? Models	7	2881	July 19, 2023

Why does the falcon QLoRA tutorial code use eos_token as pad_token?

Related topics