| Best Generations with | |
| ``` | |
| from transformers import OPTForCausalLM | |
| from transformers import GPT2Tokenizer | |
| model = OPTForCausalLM.from_pretrained("Kigo/OPT-350m-COVID-Finetune", from_tf=True) | |
| tokenizer = GPT2Tokenizer.from_pretrained("facebook/opt-350m") | |
| inputs = tokenizer("Covid-19 is Positive, 42.237% of Lungs show GGO, Lower Left Lobe was most affected, Upper Left Lobe was least affected, yes Vascular dilatation, found consolidation. The patient is in critical condition. \n\n\n ", return_tensors="pt") | |
| generate_ids = model.generate(inputs.input_ids, | |
| do_sample=True, | |
| max_new_tokens=1000, | |
| ) | |
| completion = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0] | |
| print(completion) | |
| ``` |