I’m trying to use whisper to generate transcription. I can get results when using model.generate() . But when I try to use out = model(inputs, decoder_input_ids = torch.tensor([[50258]]).to('cuda')) predicted_ids = torch.argmax(out.logits, dim=-1) transcription = processor.batch_decode(…

We can get hidden_states_embeddings from model.generate by enabling the two flags , output_hidden_states=True, return_dict_in_generate=True.

Decode whisper logits to transcript using forward instead of generate method

🤗Transformers

Hannan December 16, 2022, 9:12am 3

Is it possible to access logits by using model.generate?
However it is possible to get logits from model but as you mentioned, I couldn’t get more than one token.

Topic		Replies	Views
Disparity between output from `forward` and `generate` for greedy search (using Whisper) 🤗Transformers	3	1350	August 11, 2024
Using model() instead of model.generate() 🤗Transformers	3	519	January 30, 2025
How can I obtain the logits via model.generate()? 🤗Transformers	2	3018	October 8, 2024
How to generate without decoding? 🤗Transformers	1	371	December 13, 2023
Can I get logits for each sequence I acqired from model.generate()? Beginners	1	1313	November 27, 2020

Decode whisper logits to transcript using forward instead of generate method

Related topics