--- base_model: qresearch/doubutsu-2b-pt-756 library_name: peft license: apache-2.0 datasets: - abhishek/vqa_small --- # doubutsu-2b-lora-756-vqa An adapter for [qresearch/doubutsu-2b-pt-756](https://huggingface.co/qresearch/doubutsu-2b-pt-756) trained on [vqa_small](https://huggingface.co/datasets/abhishek/vqa_small) for 3 epochs. ## Usage ```python import torch from transformers import AutoModelForCausalLM, AutoTokenizer from PIL import Image model_id = "qresearch/doubutsu-2b-pt-756" model = AutoModelForCausalLM.from_pretrained( model_id, trust_remote_code=True, torch_dtype=torch.float16, ).to("cuda") tokenizer = AutoTokenizer.from_pretrained( model_id, use_fast=True, ) model.load_adapter("qresearch/doubutsu-2b-lora-756-vqa") image = Image.open("IMAGE") print( model.answer_question( image, "Describe the image", tokenizer, max_new_tokens=128, temperature=0.1 ), ) ``` ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - batch_size: 2 - gradient_accumulation_steps: 4 - total_train_batch_size: 8 - num_epochs: 3 ``` .x+=:. z` ^% .uef^" .u . . '88" <888'888k 888E~?888L I888 9888 4888> ' d888 '88%" 8888N=*8888 d888 '88%" 9888 9888 4888> ' 9888 'Y" 888E 888E I888 9888 4888> 8888.+" %8" R88 8888.+" 9888 9888 4888> 9888 888E 888E I888 9888 .d888L .+ 8888L @8Wou 9% 8888L 9888 9888 .d888L .+ 9888 888E 888E `888Nx?888 ^"8888*" '8888c. .+ .888888P` '8888c. .+ 9888 9888 ^"8888*" ?8888u../ 888E 888E "88" '888 "Y" "88888% ` ^"F "88888% "888*""888" "Y" "8888P' m888N= 888> 88E "YP' "YP' ^Y" ^Y' "P' `Y" 888 98> J88" '8 @% ` :" ```