├── .dockerignore ├── .editorconfig ├── .gitattributes ├── .gitignore ├── LICENSE ├── README.md ├── cog.yaml ├── data_generation ├── README.md ├── error_injection.py ├── image_weakening_utils │ ├── add_noise.py │ └── sample.py └── run_llava_image_weakening.py ├── docs ├── Customize_Component.md ├── Data.md ├── Evaluation.md ├── Finetune_Custom_Data.md ├── Intel.md ├── LLaVA_Bench.md ├── LLaVA_from_LLaMA2.md ├── LoRA.md ├── MODEL_ZOO.md ├── ScienceQA.md ├── Windows.md └── macOS.md ├── images ├── demo_cli.gif ├── happy.jpg ├── llava_example_cmp.png ├── llava_logo.png ├── llava_v1_5_radar.jpg └── logo.png ├── llava ├── __init__.py ├── constants.py ├── conversation.py ├── eval │ ├── bootstrap.py │ ├── bootstrap_dpo.py │ ├── eval_gpt_review.py │ ├── eval_gpt_review_bench.py │ ├── eval_gpt_review_visual.py │ ├── eval_multi.sh │ ├── eval_multi_safeguard.sh │ ├── eval_pope.py │ ├── eval_science_qa.py │ ├── eval_science_qa_gpt4.py │ ├── eval_science_qa_gpt4_requery.py │ ├── eval_textvqa.py │ ├── generate_webpage_data_from_table.py │ ├── m4c_evaluator.py │ ├── model_check_logits.py │ ├── model_qa.py │ ├── model_vqa.py │ ├── model_vqa_loader.py │ ├── model_vqa_mathbench.py │ ├── model_vqa_mmbench.py │ ├── model_vqa_mmvet.py │ ├── model_vqa_qbench.py │ ├── model_vqa_science.py │ ├── qa_baseline_gpt35.py │ ├── reward_score.py │ ├── robustness_eval.py │ ├── run_llava.py │ ├── summarize_gpt_review.py │ └── webpage │ │ ├── figures │ │ ├── alpaca.png │ │ ├── bard.jpg │ │ ├── chatgpt.svg │ │ ├── llama.jpg │ │ ├── swords_FILL0_wght300_GRAD0_opsz48.svg │ │ └── vicuna.jpeg │ │ ├── index.html │ │ ├── script.js │ │ └── styles.css ├── mm_utils.py ├── model │ ├── __init__.py │ ├── apply_delta.py │ ├── builder.py │ ├── consolidate.py │ ├── language_model │ │ ├── llava_llama.py │ │ ├── llava_mpt.py │ │ └── mpt │ │ │ ├── adapt_tokenizer.py │ │ │ ├── attention.py │ │ │ ├── blocks.py │ │ │ ├── configuration_mpt.py │ │ │ ├── custom_embedding.py │ │ │ ├── flash_attn_triton.py │ │ │ ├── hf_prefixlm_converter.py │ │ │ ├── meta_init_context.py │ │ │ ├── modeling_mpt.py │ │ │ ├── norm.py │ │ │ └── param_init_fns.py │ ├── llava_arch.py │ ├── make_delta.py │ ├── multimodal_encoder │ │ ├── builder.py │ │ └── clip_encoder.py │ ├── multimodal_projector │ │ └── builder.py │ └── utils.py ├── serve │ ├── __init__.py │ ├── cli.py │ ├── cli_adap.py │ ├── cli_polite.py │ ├── cli_reward.py │ ├── controller.py │ ├── examples │ │ ├── extreme_ironing.jpg │ │ └── waterview.jpg │ ├── gradio_web_server.py │ ├── model_worker.py │ ├── register_worker.py │ └── test_message.py ├── train │ ├── bpo_llava.py │ ├── bpo_llava_flash.py │ ├── llama_flash_attn_monkey_patch.py │ ├── llama_xformers_attn_monkey_patch.py │ ├── llava_trainer.py │ ├── train.py │ ├── train_mem.py │ └── train_xformers.py └── utils.py ├── main.py ├── playground └── data │ └── eval │ └── mmvet_images │ ├── v1_0.png │ ├── v1_1.png │ ├── v1_100.png │ ├── v1_101.png │ ├── v1_102.png │ ├── v1_103.png │ ├── v1_104.png │ ├── v1_105.png │ ├── v1_106.jpg │ ├── v1_107.jpg │ ├── v1_108.jpg │ ├── v1_109.jpg │ ├── v1_11.jpg │ ├── v1_110.jpg │ ├── v1_111.jpg │ ├── v1_112.jpg │ ├── v1_113.jpg │ ├── v1_114.jpg │ ├── v1_115.jpg │ ├── v1_116.jpg │ ├── v1_117.jpg │ ├── v1_118.jpg │ ├── v1_119.jpg │ ├── v1_120.jpg │ ├── v1_121.jpg │ ├── v1_122.jpg │ ├── v1_123.jpg │ ├── v1_124.jpg │ ├── v1_125.jpg │ ├── v1_126.jpg │ ├── v1_127.jpg │ ├── v1_128.jpg │ ├── v1_129.jpg │ ├── v1_13.jpg │ ├── v1_130.jpg │ ├── v1_131.jpg │ ├── v1_132.jpg │ ├── v1_133.jpg │ ├── v1_134.jpg │ ├── v1_135.jpg │ ├── v1_136.jpg │ ├── v1_137.jpg │ ├── v1_138.jpg │ ├── v1_139.jpg │ ├── v1_140.jpg │ ├── v1_141.jpg │ ├── v1_142.jpg │ ├── v1_143.jpg │ ├── v1_144.jpg │ ├── v1_145.jpg │ ├── v1_146.jpg │ ├── v1_147.jpg │ ├── v1_148.jpg │ ├── v1_149.jpg │ ├── v1_15.png │ ├── v1_150.jpg │ ├── v1_151.jpg │ ├── v1_152.jpg │ ├── v1_153.jpg │ ├── v1_154.png │ ├── v1_155.png │ ├── v1_156.jpg │ ├── v1_157.jpg │ ├── v1_158.jpg │ ├── v1_159.jpg │ ├── v1_16.jpg │ ├── v1_160.jpg │ ├── v1_161.jpg │ ├── v1_162.jpg │ ├── v1_163.jpg │ ├── v1_164.jpg │ ├── v1_165.jpg │ ├── v1_166.jpg │ ├── v1_167.jpg │ ├── v1_168.jpg │ ├── v1_169.jpg │ ├── v1_17.jpg │ ├── v1_170.jpg │ ├── v1_171.jpg │ ├── v1_172.jpg │ ├── v1_173.jpg │ ├── v1_174.jpg │ ├── v1_175.jpg │ ├── v1_176.jpg │ ├── v1_177.jpg │ ├── v1_178.jpg │ ├── v1_179.jpg │ ├── v1_18.jpg │ ├── v1_180.jpg │ ├── v1_181.jpg │ ├── v1_182.jpg │ ├── v1_183.jpg │ ├── v1_184.jpg │ ├── v1_185.jpg │ ├── v1_186.jpg │ ├── v1_187.jpg │ ├── v1_188.jpg │ ├── v1_189.jpg │ ├── v1_190.jpg │ ├── v1_191.jpg │ ├── v1_192.jpg │ ├── v1_193.jpg │ ├── v1_194.jpg │ ├── v1_195.jpg │ ├── v1_196.jpg │ ├── v1_197.jpg │ ├── v1_198.jpg │ ├── v1_199.jpg │ ├── v1_20.jpg │ ├── v1_200.jpg │ ├── v1_201.jpg │ ├── v1_202.jpg │ ├── v1_203.jpg │ ├── v1_204.jpg │ ├── v1_205.jpg │ ├── v1_206.jpg │ ├── v1_207.jpg │ ├── v1_208.png │ ├── v1_209.png │ ├── v1_21.jpg │ ├── v1_210.png │ ├── v1_211.png │ ├── v1_212.png │ ├── v1_213.jpg │ ├── v1_214.png │ ├── v1_215.jpg │ ├── v1_216.png │ ├── v1_217.png │ ├── v1_23.png │ ├── v1_25.png │ ├── v1_27.png │ ├── v1_28.jpg │ ├── v1_3.png │ ├── v1_30.jpg │ ├── v1_31.png │ ├── v1_32.png │ ├── v1_33.png │ ├── v1_34.jpg │ ├── v1_35.png │ ├── v1_36.jpg │ ├── v1_37.jpg │ ├── v1_38.jpg │ ├── v1_39.jpg │ ├── v1_40.png │ ├── v1_41.png │ ├── v1_42.png │ ├── v1_44.png │ ├── v1_46.png │ ├── v1_48.jpg │ ├── v1_49.jpg │ ├── v1_5.png │ ├── v1_50.jpg │ ├── v1_52.png │ ├── v1_53.png │ ├── v1_54.png │ ├── v1_55.png │ ├── v1_56.png │ ├── v1_57.jpg │ ├── v1_58.png │ ├── v1_60.png │ ├── v1_62.png │ ├── v1_64.png │ ├── v1_65.png │ ├── v1_66.jpg │ ├── v1_67.png │ ├── v1_68.png │ ├── v1_69.jpg │ ├── v1_7.png │ ├── v1_70.png │ ├── v1_71.jpg │ ├── v1_72.png │ ├── v1_73.jpg │ ├── v1_74.jpg │ ├── v1_75.jpg │ ├── v1_76.png │ ├── v1_77.jpg │ ├── v1_78.jpg │ ├── v1_79.jpg │ ├── v1_8.png │ ├── v1_80.jpg │ ├── v1_81.jpg │ ├── v1_82.png │ ├── v1_83.png │ ├── v1_84.png │ ├── v1_85.jpg │ ├── v1_86.jpg │ ├── v1_87.jpg │ ├── v1_88.jpg │ ├── v1_89.jpg │ ├── v1_9.png │ ├── v1_90.jpg │ ├── v1_91.jpg │ ├── v1_92.jpg │ ├── v1_93.jpg │ ├── v1_94.jpg │ ├── v1_95.jpg │ ├── v1_96.jpg │ ├── v1_97.jpg │ ├── v1_98.jpg │ └── v1_99.png ├── predict.py ├── pyproject.toml ├── qwen ├── README.md ├── dpo_config │ └── example.yaml ├── launch_dpo.py ├── requirements.txt └── run_dpo.py └── scripts ├── concatenate_json.py ├── convert_gqa_for_eval.py ├── convert_mmbench_for_submission.py ├── convert_mmvet_for_eval.py ├── convert_seed_for_submission.py ├── convert_sqa_to_llava.py ├── convert_sqa_to_llava_base_prompt.py ├── convert_vizwiz_for_submission.py ├── convert_vqav2_for_submission.py ├── eval_mmvet.h ├── extract_mm_projector.py ├── finetune.sh ├── finetune_bpo.sh ├── finetune_bpo_flash.sh ├── finetune_full_schedule.sh ├── finetune_lora.sh ├── finetune_qlora.sh ├── finetune_sft.sh ├── finetune_sqa.sh ├── merge_lora_weights.py ├── merge_lora_weights_reward.py ├── pretrain.sh ├── pretrain_xformers.sh ├── sqa_eval_batch.sh ├── sqa_eval_gather.sh └── v1_5 ├── eval ├── bootstrap_dpo_multi.sh ├── bootstrap_multi.sh ├── eval_multi.sh ├── eval_multi_lora.sh ├── eval_multi_math.sh ├── eval_multi_pope.sh ├── eval_multi_pope_full.sh ├── gqa.sh ├── llavabench.sh ├── mmbench.sh ├── mmbench_cn.sh ├── mme.sh ├── mmvet.sh ├── pope.sh ├── qbench.sh ├── qbench_zh.sh ├── score_multi.sh ├── seed.sh ├── sqa.sh ├── textvqa.sh ├── vizwiz.sh └── vqav2.sh ├── finetune.sh ├── finetune_lora.sh ├── finetune_task.sh ├── finetune_task_lora.sh └── pretrain.sh /.dockerignore: -------------------------------------------------------------------------------- 1 | # The .dockerignore file excludes files from the container build process. 2 | # 3 | # https://docs.docker.com/engine/reference/builder/#dockerignore-file 4 | 5 | # Exclude Git files 6 | .git 7 | .github 8 | .gitignore 9 | 10 | # Exclude Python cache files 11 | __pycache__ 12 | .mypy_cache 13 | .pytest_cache 14 | .ruff_cache 15 | 16 | # Exclude Python virtual environment 17 | /venv 18 | 19 | # Exclude some weights 20 | /openai 21 | /liuhaotian 22 | -------------------------------------------------------------------------------- /.editorconfig: -------------------------------------------------------------------------------- 1 | root = true 2 | 3 | # Unix-style newlines with a newline ending every file 4 | [*] 5 | end_of_line = lf 6 | insert_final_newline = true 7 | trim_trailing_whitespace = true 8 | charset = utf-8 9 | 10 | # 4 space indentation 11 | [*.{py,json}] 12 | indent_style = space 13 | indent_size = 4 14 | 15 | # 2 space indentation 16 | [*.{md,sh,yaml,yml}] 17 | indent_style = space 18 | indent_size = 2 -------------------------------------------------------------------------------- /.gitattributes: -------------------------------------------------------------------------------- 1 | # https://git-scm.com/docs/gitattributes 2 | 3 | # Set the default behavior, in case people don't have core.autocrlf set. 4 | # https://git-scm.com/docs/gitattributes#_end_of_line_conversion 5 | * text=auto 6 | 7 | # common python attributes, taken from https://github.com/alexkaratarakis/gitattributes/blob/710900479a2bedeec7003d381719521ffbb18bf8/Python.gitattributes 8 | # Source files 9 | # ============ 10 | *.pxd text diff=python 11 | *.py text diff=python 12 | *.py3 text diff=python 13 | *.pyw text diff=python 14 | *.pyx text diff=python 15 | *.pyz text diff=python 16 | *.pyi text diff=python 17 | 18 | # Binary files 19 | # ============ 20 | *.db binary 21 | *.p binary 22 | *.pkl binary 23 | *.pickle binary 24 | *.pyc binary export-ignore 25 | *.pyo binary export-ignore 26 | *.pyd binary 27 | 28 | # Jupyter notebook 29 | *.ipynb text eol=lf 30 | -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | # Python 2 | __pycache__ 3 | *.pyc 4 | *.egg-info 5 | dist 6 | 7 | # Log 8 | *.log 9 | *.log.* 10 | *.json 11 | *.jsonl 12 | 13 | # Data 14 | !**/alpaca-data-conversation.json 15 | 16 | # Editor 17 | .idea 18 | *.swp 19 | 20 | # Other 21 | .DS_Store 22 | wandb 23 | output 24 | 25 | checkpoints 26 | ckpts* 27 | 28 | .ipynb_checkpoints 29 | *.ipynb 30 | 31 | # DevContainer 32 | !.devcontainer/* 33 | 34 | # Demo 35 | serve_images/ 36 | 37 | # debug 38 | debug/ 39 | results/ 40 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 |
2 | BPO 3 |

Generated by DALL·E 3

4 |
5 | 6 | This repository contains the code for the paper titled "Strengthening Multimodal Large Language Model with Bootstrapped Preference Optimization". [[Link to our paper](https://arxiv.org/abs/2403.08730)] 7 | ## Install Packages 8 | 9 | ``` 10 | 11 | conda create -n bpo python=3.10 -y 12 | 13 | conda activate bpo 14 | 15 | pip install -e . 16 | 17 | ``` 18 | Install flash attention for efficient training 19 | 20 | ``` 21 | pip install -e ".[train]" 22 | 23 | pip install flash-attn --no-build-isolation 24 | ``` 25 | ## Training data 26 | Download ShareGPT4V from [here](https://huggingface.co/datasets/Lin-Chen/ShareGPT4V) 27 | 28 | Download COCO from [here](https://cocodataset.org/#home) 29 | 30 | Download dataset annotation from [here](https://huggingface.co/datasets/renjiepi/BPO_Instruct) 31 | 32 | Extract data from ShareGPT4V and organize the images as follows: 33 | 34 | ``` 35 | Image_root 36 | ├── coco/ 37 | ├──train2017/ 38 | ├── llava/ 39 | ├──llava_pretrain/ 40 | ├── sam/ 41 | ├── share_textvqa/ 42 | ├──images/ 43 | ├── web-celebrity/ 44 | ├──images/ 45 | ├── web-landmark/ 46 | ├──images/ 47 | ├── wikiart/ 48 | ├──images/ 49 | ``` 50 | 51 | ## Training 52 | ### Training BPO 53 | ``` 54 | bash scripts/finetune_bpo.sh 55 | ``` 56 | ### Training BPO with flash attention 57 | ``` 58 | bash scripts/finetune_bpo_flash.sh 59 | ``` 60 | ## Acknowledgement 61 | The project is built on top of the amazing multimodal large language model [LLaVA](https://github.com/haotian-liu/LLaVA), RLHF package [trl](https://github.com/huggingface/trl), DPO for multimodal learning [Silkie](https://github.com/vlf-silkie/VLFeedback), and visual contrastive decoding [VCD](https://github.com/DAMO-NLP-SG/VCD). 62 | Thanks for these great work! 63 | 64 | 65 | If you find our work useful for your research or applications, please cite using this BibTeX: 66 | ```bibtex 67 | @misc{pi2024strengthening, 68 | title={Strengthening Multimodal Large Language Model with Bootstrapped Preference Optimization}, 69 | author={Renjie Pi and Tianyang Han and Wei Xiong and Jipeng Zhang and Runtao Liu and Rui Pan and Tong Zhang}, 70 | year={2024}, 71 | eprint={2403.08730}, 72 | archivePrefix={arXiv}, 73 | primaryClass={cs.CL} 74 | } 75 | ``` 76 | -------------------------------------------------------------------------------- /cog.yaml: -------------------------------------------------------------------------------- 1 | # Configuration for Cog ⚙️ 2 | # Reference: https://github.com/replicate/cog/blob/main/docs/yaml.md 3 | 4 | build: 5 | gpu: true 6 | 7 | python_version: "3.11" 8 | 9 | python_packages: 10 | - "torch==2.0.1" 11 | - "accelerate==0.21.0" 12 | - "bitsandbytes==0.41.0" 13 | - "deepspeed==0.9.5" 14 | - "einops-exts==0.0.4" 15 | - "einops==0.6.1" 16 | - "gradio==3.35.2" 17 | - "gradio_client==0.2.9" 18 | - "httpx==0.24.0" 19 | - "markdown2==2.4.10" 20 | - "numpy==1.26.0" 21 | - "peft==0.4.0" 22 | - "scikit-learn==1.2.2" 23 | - "sentencepiece==0.1.99" 24 | - "shortuuid==1.0.11" 25 | - "timm==0.6.13" 26 | - "tokenizers==0.13.3" 27 | - "torch==2.0.1" 28 | - "torchvision==0.15.2" 29 | - "transformers==4.31.0" 30 | - "wandb==0.15.12" 31 | - "wavedrom==2.0.3.post3" 32 | - "Pygments==2.16.1" 33 | run: 34 | - curl -o /usr/local/bin/pget -L "https://github.com/replicate/pget/releases/download/v0.0.3/pget" && chmod +x /usr/local/bin/pget 35 | 36 | # predict.py defines how predictions are run on your model 37 | predict: "predict.py:Predictor" 38 | -------------------------------------------------------------------------------- /data_generation/README.md: -------------------------------------------------------------------------------- 1 | # Data Generation for BPO 2 | - For image weakening, we are inspired by [VCD](https://github.com/DAMO-NLP-SG/VCD) to add noise to image features, which generates negative responses with pretraining bias. 3 | - For error injection, we utilize the pretrained LLM that is the same as the base model of the MLLM to directly inject erroneous concepts. 4 | 5 | 6 | ## Install environments 7 | For error injection, please install VLLM to speed up inference. 8 | ``` 9 | pip intall vllm 10 | ``` 11 | If it can't run, please consider building the package from source. 12 | 13 | ## Image weakening 14 | ```bash 15 | cd data_generation 16 | python run_llava_image_weakening.py --model-path liuhaotian/llava-v1.5-13b --image_file YOUR_IMAGE_PATH --query YOUR_JSON_PATH --save_path OUTPUT_PATH 17 | ``` 18 | 19 | `YOUR_JSON_PATH` should be a list of json file and has the following format 20 | ``` 21 | [ 22 | {'prompt': 'What do you see happening in this image?\n', 23 | 'image': 'coco/train2017/000000000009.jpg', 24 | 'completions': [{'score': 1, 25 | 'response': "xxxx", 26 | 'type': 'gt'}]}, 27 | 28 | {'prompt': 'question\n', 29 | 'image': 'coco/train2017/0000000000010.jpg', 30 | 'completions': [{'score': 1, 31 | 'response': "xxxx", 32 | 'type': 'gt'}]}, 33 | ....... 34 | ] 35 | ``` 36 | Note: Image weakening currently runs on generic inference pipeline. We will consider integrating it with MLLM acceleration framework, e.g,. https://github.com/InternLM/lmdeploy. 37 | 38 | ## Error injection 39 | ```bash 40 | cd data_generation 41 | python error_injection.py --model_name_or_path PATH-TO-LLM --dataset_path PATH-TO-SFT-DATA --output_result_path PATH-TO-PREFERENCE-DATA 42 | ``` 43 | 44 | -------------------------------------------------------------------------------- /data_generation/image_weakening_utils/add_noise.py: -------------------------------------------------------------------------------- 1 | import torch 2 | 3 | def add_diffusion_noise(image_tensor, noise_step): 4 | num_steps = 1000 # Number of diffusion steps 5 | 6 | # decide beta in each step 7 | betas = torch.linspace(-6,6,num_steps) 8 | betas = torch.sigmoid(betas) * (0.5e-2 - 1e-5) + 1e-5 9 | 10 | # decide alphas in each step 11 | alphas = 1 - betas 12 | alphas_prod = torch.cumprod(alphas, dim=0) 13 | alphas_prod_p = torch.cat([torch.tensor([1]).float(), alphas_prod[:-1]],0) # p for previous 14 | alphas_bar_sqrt = torch.sqrt(alphas_prod) 15 | one_minus_alphas_bar_log = torch.log(1 - alphas_prod) 16 | one_minus_alphas_bar_sqrt = torch.sqrt(1 - alphas_prod) 17 | 18 | def q_x(x_0,t): 19 | noise = torch.randn_like(x_0) 20 | alphas_t = alphas_bar_sqrt[t] 21 | alphas_1_m_t = one_minus_alphas_bar_sqrt[t] 22 | return (alphas_t*x_0 + alphas_1_m_t*noise) 23 | 24 | noise_delta = int(noise_step) # from 0-999 25 | noisy_image = image_tensor.clone() 26 | image_tensor_cd = q_x(noisy_image,noise_step) 27 | 28 | return image_tensor_cd 29 | 30 | -------------------------------------------------------------------------------- /docs/Customize_Component.md: -------------------------------------------------------------------------------- 1 | # Customize Components in LLaVA 2 | 3 | This is an initial guide on how to replace the LLMs, visual encoders, etc. with your choice of components. 4 | 5 | ## LLM 6 | 7 | It is quite simple to swap out LLaMA to any other LLMs. You can refer to our implementation of [`llava_llama.py`](https://raw.githubusercontent.com/haotian-liu/LLaVA/main/llava/model/language_model/llava_llama.py) for an example of how to replace the LLM. 8 | 9 | Although it may seem that it still needs ~100 lines of code, most of them are copied from the original `llama.py` from HF. The only part that is different is to insert some lines for processing the multimodal inputs. 10 | 11 | In `forward` function, you can see that we call `self.prepare_inputs_labels_for_multimodal` to process the multimodal inputs. This function is defined in `LlavaMetaForCausalLM` and you just need to insert it into the `forward` function of your LLM. 12 | 13 | In `prepare_inputs_for_generation` function, you can see that we add `images` to the `model_inputs`. This is because we need to pass the images to the LLM during generation. 14 | 15 | These are basically all the changes you need to make to replace the LLM. 16 | 17 | ## Visual Encoder 18 | 19 | You can check out [`clip_encoder.py`](https://github.com/haotian-liu/LLaVA/blob/main/llava/model/multimodal_encoder/clip_encoder.py) on how we implement the CLIP visual encoder. 20 | 21 | -------------------------------------------------------------------------------- /docs/Data.md: -------------------------------------------------------------------------------- 1 | ## Data 2 | 3 | | Data file name | Size | 4 | | --- | ---: | 5 | | [llava_instruct_150k.json](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/blob/main/llava_instruct_150k.json) | 229 MB | 6 | | [llava_instruct_80k.json](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/blob/main/llava_instruct_80k.json) | 229 MB | 7 | | [conversation_58k.json](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/blob/main/conversation_58k.json) | 126 MB | 8 | | [detail_23k.json](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/blob/main/detail_23k.json) | 20.5 MB | 9 | | [complex_reasoning_77k.json](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/blob/main/complex_reasoning_77k.json) | 79.6 MB | 10 | 11 | ### Pretraining Dataset 12 | The pretraining dataset used in this release is a subset of CC-3M dataset, filtered with a more balanced concept coverage distribution. Please see [here](https://huggingface.co/datasets/liuhaotian/LLaVA-CC3M-Pretrain-595K) for a detailed description of the dataset structure and how to download the images. 13 | 14 | If you already have CC-3M dataset on your disk, the image names follow this format: `GCC_train_000000000.jpg`. You may edit the `image` field correspondingly if necessary. 15 | 16 | | Data | Chat File | Meta Data | Size | 17 | | --- | --- | --- | ---: | 18 | | CC-3M Concept-balanced 595K | [chat.json](https://huggingface.co/datasets/liuhaotian/LLaVA-CC3M-Pretrain-595K/blob/main/chat.json) | [metadata.json](https://huggingface.co/datasets/liuhaotian/LLaVA-CC3M-Pretrain-595K/blob/main/metadata.json) | 211 MB 19 | | LAION/CC/SBU BLIP-Caption Concept-balanced 558K | [blip_laion_cc_sbu_558k.json](https://huggingface.co/datasets/liuhaotian/LLaVA-Pretrain/blob/main/blip_laion_cc_sbu_558k.json) | [metadata.json](#) | 181 MB 20 | 21 | **Important notice**: Upon the request from the community, as ~15% images of the original CC-3M dataset are no longer accessible, we upload [`images.zip`](https://huggingface.co/datasets/liuhaotian/LLaVA-CC3M-Pretrain-595K/blob/main/images.zip) for better reproducing our work in research community. It must not be used for any other purposes. The use of these images must comply with the CC-3M license. This may be taken down at any time when requested by the original CC-3M dataset owner or owners of the referenced images. 22 | 23 | ### GPT-4 Prompts 24 | 25 | We provide our prompts and few-shot samples for GPT-4 queries, to better facilitate research in this domain. Please check out the [`prompts`](https://github.com/haotian-liu/LLaVA/tree/main/playground/data/prompts) folder for three kinds of questions: conversation, detail description, and complex reasoning. 26 | 27 | They are organized in a format of `system_message.txt` for system message, pairs of `abc_caps.txt` for few-shot sample user input, and `abc_conv.txt` for few-shot sample reference output. 28 | 29 | Note that you may find them in different format. For example, `conversation` is in `jsonl`, and detail description is answer-only. The selected format in our preliminary experiments works slightly better than a limited set of alternatives that we tried: `jsonl`, more natural format, answer-only. If interested, you may try other variants or conduct more careful study in this. Contributions are welcomed! 30 | -------------------------------------------------------------------------------- /docs/Finetune_Custom_Data.md: -------------------------------------------------------------------------------- 1 | # Finetune LLaVA on Custom Datasets 2 | 3 | ## Dataset Format 4 | 5 | Convert your data to a JSON file of a List of all samples. Sample metadata should contain `id` (a unique identifier), `image` (the path to the image), and `conversations` (the conversation data between human and AI). 6 | 7 | A sample JSON for finetuning LLaVA for generating tag-style captions for Stable Diffusion: 8 | 9 | ```json 10 | [ 11 | { 12 | "id": "997bb945-628d-4724-b370-b84de974a19f", 13 | "image": "part-000001/997bb945-628d-4724-b370-b84de974a19f.jpg", 14 | "conversations": [ 15 | { 16 | "from": "human", 17 | "value": "\nWrite a prompt for Stable Diffusion to generate this image." 18 | }, 19 | { 20 | "from": "gpt", 21 | "value": "a beautiful painting of chernobyl by nekro, pascal blanche, john harris, greg rutkowski, sin jong hun, moebius, simon stalenhag. in style of cg art. ray tracing. cel shading. hyper detailed. realistic. ue 5. maya. octane render. " 22 | }, 23 | ] 24 | }, 25 | ... 26 | ] 27 | ``` 28 | 29 | ## Command 30 | 31 | If you have a limited task-specific data, we recommend finetuning from LLaVA checkpoints with LoRA following this [script](https://github.com/haotian-liu/LLaVA/blob/main/scripts/v1_5/finetune_task_lora.sh). 32 | 33 | If the amount of the task-specific data is sufficient, you can also finetune from LLaVA checkpoints with full-model finetuning following this [script](https://github.com/haotian-liu/LLaVA/blob/main/scripts/v1_5/finetune_task.sh). 34 | 35 | You may need to adjust the hyperparameters to fit each specific dataset and your hardware constraint. 36 | 37 | 38 | -------------------------------------------------------------------------------- /docs/Intel.md: -------------------------------------------------------------------------------- 1 | # Intel Platforms 2 | 3 | * Support [Intel GPU Max Series](https://www.intel.com/content/www/us/en/products/details/discrete-gpus/data-center-gpu/max-series.html) 4 | * Support [Intel CPU Sapphire Rapides](https://ark.intel.com/content/www/us/en/ark/products/codename/126212/products-formerly-sapphire-rapids.html) 5 | * Based on [Intel Extension for Pytorch](https://intel.github.io/intel-extension-for-pytorch) 6 | 7 | More details in [**intel branch**](https://github.com/haotian-liu/LLaVA/tree/intel/docs/intel) 8 | -------------------------------------------------------------------------------- /docs/LLaVA_from_LLaMA2.md: -------------------------------------------------------------------------------- 1 | # LLaVA (based on Llama 2 LLM, Preview) 2 | 3 | *NOTE: This is a technical preview. We are still running hyperparameter search, and will release the final model soon. If you'd like to contribute to this, please contact us.* 4 | 5 | :llama: **-Introduction-** [Llama 2 is an open-source LLM released by Meta AI](https://about.fb.com/news/2023/07/llama-2/) today (July 18, 2023). Compared with its early version [Llama 1](https://ai.meta.com/blog/large-language-model-llama-meta-ai/), Llama 2 is more favored in ***stronger language performance***, ***longer context window***, and importantly ***commercially usable***! While Llama 2 is changing the LLM market landscape in the language space, its multimodal ability remains unknown. We quickly develop the LLaVA variant based on the latest Llama 2 checkpoints, and release it to the community for the public use. 6 | 7 | You need to apply for and download the latest Llama 2 checkpoints to start your own training (apply [here](https://ai.meta.com/resources/models-and-libraries/llama-downloads/)) 8 | 9 | 10 | ## Training 11 | 12 | Please checkout [`pretrain.sh`](https://github.com/haotian-liu/LLaVA/blob/main/scripts/pretrain.sh), [`finetune.sh`](https://github.com/haotian-liu/LLaVA/blob/main/scripts/finetune.sh), [`finetune_lora.sh`](https://github.com/haotian-liu/LLaVA/blob/main/scripts/finetune_lora.sh). 13 | 14 | ## LLaVA (based on Llama 2), What is different? 15 | 16 | :volcano: How is the new LLaVA based on Llama 2 different from Llama 1? The comparisons of the training process are described: 17 | - **Pre-training**. The pre-trained base LLM is changed from Llama 1 to Llama 2 18 | - **Language instruction-tuning**. The previous LLaVA model starts with Vicuna, which is instruct tuned on ShareGPT data from Llama 1; The new LLaVA model starts with Llama 2 Chat, which is an instruct tuned checkpoint on dialogue data from Llama 2. 19 | - **Multimodal instruction-tuning**. The same LLaVA-Lighting process is applied. 20 | 21 | 22 | ### Results 23 | 24 | - Llama 2 is better at following the instructions of role playing; Llama 2 fails in following the instructions of translation 25 | - The quantitative evaluation on [LLaVA-Bench](https://github.com/haotian-liu/LLaVA/blob/main/docs/LLaVA_Bench.md) demonstrates on-par performance between Llama 2 and Llama 1 in LLaVA's multimodal chat ability. 26 | 27 | 28 | 29 | 30 | -------------------------------------------------------------------------------- /docs/ScienceQA.md: -------------------------------------------------------------------------------- 1 | ### ScienceQA 2 | 3 | #### Prepare Data 4 | 1. Please see ScienceQA [repo](https://github.com/lupantech/ScienceQA) for setting up the dataset. 5 | 2. Generate ScienceQA dataset for LLaVA conversation-style format. 6 | 7 | ```Shell 8 | python scripts/convert_sqa_to_llava.py \ 9 | convert_to_llava \ 10 | --base-dir /path/to/ScienceQA/data/scienceqa \ 11 | --prompt-format "QCM-LEA" \ 12 | --split {train,val,minival,test,minitest} 13 | ``` 14 | 15 | #### Training 16 | 17 | 1. Pretraining 18 | 19 | You can download our pretrained projector weights from our [Model Zoo](), or train your own projector weights using [`pretrain.sh`](https://github.com/haotian-liu/LLaVA/blob/main/scripts/pretrain.sh). 20 | 21 | 2. Finetuning 22 | 23 | See [`finetune_sqa.sh`](https://github.com/haotian-liu/LLaVA/blob/main/scripts/finetune_sqa.sh). 24 | 25 | #### Evaluation 26 | 27 | 1. Multiple-GPU inference 28 | You may evaluate this with multiple GPUs, and concatenate the generated jsonl files. Please refer to our script for [batch evaluation](https://github.com/haotian-liu/LLaVA/blob/main/scripts/sqa_eval_batch.sh) and [results gathering](https://github.com/haotian-liu/LLaVA/blob/main/scripts/sqa_eval_gather.sh). 29 | 30 | 2. Single-GPU inference 31 | 32 | (a) Generate LLaVA responses on ScienceQA dataset 33 | 34 | ```Shell 35 | python -m llava.eval.model_vqa_science \ 36 | --model-path liuhaotian/llava-lcs558k-scienceqa-vicuna-13b-v1.3 \ 37 | --question-file /path/to/ScienceQA/data/scienceqa/llava_test_QCM-LEA.json \ 38 | --image-folder /path/to/ScienceQA/data/scienceqa/images/test \ 39 | --answers-file vqa/results/ScienceQA/test_llava-13b.jsonl \ 40 | --conv-mode llava_v1 41 | ``` 42 | 43 | (b) Evaluate the generated responses 44 | 45 | ```Shell 46 | python eval_science_qa.py \ 47 | --base-dir /path/to/ScienceQA/data/scienceqa \ 48 | --result-file vqa/results/ScienceQA/test_llava-13b.jsonl \ 49 | --output-file vqa/results/ScienceQA/test_llava-13b_output.json \ 50 | --output-result vqa/results/ScienceQA/test_llava-13b_result.json \ 51 | ``` 52 | 53 | For reference, we attach our prediction file [`test_sqa_llava_lcs_558k_sqa_12e_vicuna_v1_3_13b.json`](https://github.com/haotian-liu/LLaVA/blob/main/llava/eval/table/results/test_sqa_llava_lcs_558k_sqa_12e_vicuna_v1_3_13b.json) and [`test_sqa_llava_13b_v0.json`](https://github.com/haotian-liu/LLaVA/blob/main/llava/eval/table/results/test_sqa_llava_13b_v0.json) for comparison when reproducing our results, as well as for further analysis in detail. 54 | -------------------------------------------------------------------------------- /docs/Windows.md: -------------------------------------------------------------------------------- 1 | # Run LLaVA on Windows 2 | 3 | *NOTE: LLaVA on Windows is not fully supported. Currently we only support 16-bit inference. For a more complete support, please use [WSL2](https://learn.microsoft.com/en-us/windows/wsl/install) for now. More functionalities on Windows is to be added soon, stay tuned.* 4 | 5 | ## Installation 6 | 7 | 1. Clone this repository and navigate to LLaVA folder 8 | ```bash 9 | git clone https://github.com/haotian-liu/LLaVA.git 10 | cd LLaVA 11 | ``` 12 | 13 | 2. Install Package 14 | ```Shell 15 | conda create -n llava python=3.10 -y 16 | conda activate llava 17 | python -mpip install --upgrade pip # enable PEP 660 support 18 | pip install torch==2.0.1+cu117 torchvision==0.15.2+cu117 torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu117 19 | pip install -e . 20 | pip uninstall bitsandbytes 21 | ``` 22 | 23 | ## Run demo 24 | 25 | See instructions [here](https://github.com/haotian-liu/LLaVA#demo). 26 | 27 | Note that quantization (4-bit, 8-bit) is *NOT* supported on Windows. Stay tuned for the 4-bit support on Windows! 28 | -------------------------------------------------------------------------------- /docs/macOS.md: -------------------------------------------------------------------------------- 1 | # Run LLaVA on macOS 2 | 3 | *NOTE: LLaVA on macOS is not fully supported. Currently we only support 16-bit inference. More functionalities on macOS is to be added soon, stay tuned.* 4 | 5 | ## Installation 6 | 7 | 1. Clone this repository and navigate to LLaVA folder 8 | ```bash 9 | git clone https://github.com/haotian-liu/LLaVA.git 10 | cd LLaVA 11 | ``` 12 | 13 | 2. Install Package 14 | ```Shell 15 | conda create -n llava python=3.10 -y 16 | conda activate llava 17 | python -mpip install --upgrade pip # enable PEP 660 support 18 | pip install -e . 19 | pip install torch==2.1.0 torchvision==0.16.0 20 | pip uninstall bitsandbytes 21 | ``` 22 | 23 | ## Run demo 24 | 25 | Specify `--device mps` when launching model worker or CLI. 26 | 27 | See instructions [here](https://github.com/haotian-liu/LLaVA#demo). 28 | 29 | Note that quantization (4-bit, 8-bit) is *NOT* supported on macOS. Stay tuned for the 4-bit support on macOS! 30 | -------------------------------------------------------------------------------- /images/demo_cli.gif: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/images/demo_cli.gif -------------------------------------------------------------------------------- /images/happy.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/images/happy.jpg -------------------------------------------------------------------------------- /images/llava_example_cmp.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/images/llava_example_cmp.png -------------------------------------------------------------------------------- /images/llava_logo.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/images/llava_logo.png -------------------------------------------------------------------------------- /images/llava_v1_5_radar.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/images/llava_v1_5_radar.jpg -------------------------------------------------------------------------------- /images/logo.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/images/logo.png -------------------------------------------------------------------------------- /llava/__init__.py: -------------------------------------------------------------------------------- 1 | from .model import LlavaLlamaForCausalLM, LlavaLlamaBPOForCausalLM 2 | -------------------------------------------------------------------------------- /llava/constants.py: -------------------------------------------------------------------------------- 1 | CONTROLLER_HEART_BEAT_EXPIRATION = 30 2 | WORKER_HEART_BEAT_INTERVAL = 15 3 | 4 | LOGDIR = "." 5 | 6 | # Model Constants 7 | IGNORE_INDEX = -100 8 | IMAGE_TOKEN_INDEX = -200 9 | DEFAULT_IMAGE_TOKEN = "" 10 | DEFAULT_IMAGE_PATCH_TOKEN = "" 11 | DEFAULT_IM_START_TOKEN = "" 12 | DEFAULT_IM_END_TOKEN = "" 13 | IMAGE_PLACEHOLDER = "" 14 | -------------------------------------------------------------------------------- /llava/eval/eval_multi.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # Check if four arguments are passed 4 | if [ "$#" -ne 4 ]; then 5 | echo "Usage: $0 " 6 | exit 1 7 | fi 8 | 9 | scenes=("01-Illegal_Activitiy" "02-HateSpeech" "03-Malware_Generation" "04-Physical_Harm" "05-EconomicHarm" "06-Fraud" "07-Sex" "08-Political_Lobbying" "09-Privacy_Violence" "10-Legal_Opinion" "11-Financial_Advice" "12-Health_Consultation" "13-Gov_Decision") 10 | 11 | # Assign the command line arguments to variables 12 | model_path=$1 13 | answers_root=$2 14 | N=$3 15 | temperature=$4 16 | 17 | # Check if the answers_root directory exists 18 | if [ ! -d "$answers_root" ]; then 19 | # Directory does not exist, so create it 20 | mkdir "$answers_root" 21 | fi 22 | 23 | for scene in "${scenes[@]}"; do 24 | answer_scene_path="${answers_root}/${scene}" 25 | if [ ! -d "$answer_scene_path" ]; then 26 | # Directory does not exist, so create it 27 | mkdir "$answer_scene_path" 28 | fi 29 | 30 | # Loop over each chunk/process 31 | for ((chunk_id = 0; chunk_id < N; chunk_id++)); do 32 | # Define the answer path for each chunk 33 | answer_path="${answer_scene_path}/${chunk_id}.json" 34 | if [ -f "$answer_path" ]; then 35 | rm "$answer_path" 36 | fi 37 | 38 | # Run the Python program in the background 39 | CUDA_VISIBLE_DEVICES="$chunk_id" python llava/eval/robustness_eval.py --model-path "$model_path" --scene "$scene" --answers_file "$answer_path" --num-chunks "$N" --chunk-idx "$chunk_id" --temperature "$temperature" & 40 | 41 | # Uncomment below if you need a slight delay between starting each process 42 | # sleep 0.1 43 | done 44 | 45 | # Wait for all background processes to finish 46 | wait 47 | cd $answer_scene_path 48 | merged_file="merged.json" 49 | if [ -f "$merged_file" ]; then 50 | rm "$merged_file" 51 | fi 52 | 53 | # Merge all the JSON files into one 54 | python ~/polite_llava/scripts/concatenate_json.py *.json 55 | cd ~/polite_llava 56 | # Remove the unmerged files 57 | for ((chunk_id = 0; chunk_id < N; chunk_id++)); do 58 | answer_path="${answer_scene_path}/${chunk_id}.json" 59 | if [ -f "$answer_path" ]; then 60 | rm "$answer_path" 61 | fi 62 | done 63 | done 64 | -------------------------------------------------------------------------------- /llava/eval/eval_multi_safeguard.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # Check if four arguments are passed 4 | if [ "$#" -ne 6 ]; then 5 | echo "Usage: $0 " 6 | exit 1 7 | fi 8 | 9 | #scenes=("01-Illegal_Activitiy" "02-HateSpeech" "03-Malware_Generation" "04-Physical_Harm" "05-EconomicHarm" "06-Fraud" "07-Sex" "08-Political_Lobbying" "09-Privacy_Violence" "10-Legal_Opinion" "11-Financial_Advice" "12-Health_Consultation" "13-Gov_Decision") 10 | scenes=("04-Physical_Harm" "05-EconomicHarm" "06-Fraud" "07-Sex" "08-Political_Lobbying" "09-Privacy_Violence" "10-Legal_Opinion" "11-Financial_Advice" "12-Health_Consultation" "13-Gov_Decision") 11 | #scenes=("07-Sex" "08-Political_Lobbying" "09-Privacy_Violence" "10-Legal_Opinion" "11-Financial_Advice" "12-Health_Consultation" "13-Gov_Decision") 12 | 13 | # Assign the command line arguments to variables 14 | model_path=$1 15 | answers_root=$2 16 | N=$3 17 | temperature=$4 18 | harm_detector=$5 19 | detoxifier=$6 20 | 21 | # Check if the answers_root directory exists 22 | if [ ! -d "$answers_root" ]; then 23 | # Directory does not exist, so create it 24 | mkdir "$answers_root" 25 | fi 26 | 27 | for scene in "${scenes[@]}"; do 28 | answer_scene_path="${answers_root}/${scene}" 29 | if [ ! -d "$answer_scene_path" ]; then 30 | # Directory does not exist, so create it 31 | mkdir "$answer_scene_path" 32 | fi 33 | 34 | # Loop over each chunk/process 35 | for ((chunk_id = 0; chunk_id < N; chunk_id++)); do 36 | # Define the answer path for each chunk 37 | answer_path="${answer_scene_path}/${chunk_id}.json" 38 | if [ -f "$answer_path" ]; then 39 | rm "$answer_path" 40 | fi 41 | 42 | # Run the Python program in the background 43 | CUDA_VISIBLE_DEVICES="$chunk_id" python llava/eval/robustness_eval.py --model-path "$model_path" --scene "$scene" --answers_file "$answer_path" --num-chunks "$N" --chunk-idx "$chunk_id" --temperature "$temperature" --harm_detector "$harm_detector" --detoxifier "$detoxifier"& 44 | 45 | # Uncomment below if you need a slight delay between starting each process 46 | # sleep 0.1 47 | done 48 | 49 | # Wait for all background processes to finish 50 | wait 51 | cd $answer_scene_path 52 | merged_file="merged.json" 53 | if [ -f "$merged_file" ]; then 54 | rm "$merged_file" 55 | fi 56 | 57 | # Merge all the JSON files into one 58 | python ~/polite_llava/scripts/concatenate_json.py *.json 59 | cd ~/polite_llava 60 | # Remove the unmerged files 61 | for ((chunk_id = 0; chunk_id < N; chunk_id++)); do 62 | answer_path="${answer_scene_path}/${chunk_id}.json" 63 | if [ -f "$answer_path" ]; then 64 | rm "$answer_path" 65 | fi 66 | done 67 | done 68 | -------------------------------------------------------------------------------- /llava/eval/eval_pope.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | import argparse 4 | 5 | def eval_pope(answers, label_file): 6 | label_list = [json.loads(q)['label'] for q in open(label_file, 'r')] 7 | 8 | for answer in answers: 9 | text = answer['text'] 10 | 11 | # Only keep the first sentence 12 | if text.find('.') != -1: 13 | text = text.split('.')[0] 14 | 15 | text = text.replace(',', '') 16 | words = text.split(' ') 17 | if 'No' in words or 'not' in words or 'no' in words: 18 | answer['text'] = 'no' 19 | else: 20 | answer['text'] = 'yes' 21 | 22 | for i in range(len(label_list)): 23 | if label_list[i] == 'no': 24 | label_list[i] = 0 25 | else: 26 | label_list[i] = 1 27 | 28 | pred_list = [] 29 | for answer in answers: 30 | if answer['text'] == 'no': 31 | pred_list.append(0) 32 | else: 33 | pred_list.append(1) 34 | 35 | pos = 1 36 | neg = 0 37 | yes_ratio = pred_list.count(1) / len(pred_list) 38 | 39 | TP, TN, FP, FN = 0, 0, 0, 0 40 | for pred, label in zip(pred_list, label_list): 41 | if pred == pos and label == pos: 42 | TP += 1 43 | elif pred == pos and label == neg: 44 | FP += 1 45 | elif pred == neg and label == neg: 46 | TN += 1 47 | elif pred == neg and label == pos: 48 | FN += 1 49 | 50 | print('TP\tFP\tTN\tFN\t') 51 | print('{}\t{}\t{}\t{}'.format(TP, FP, TN, FN)) 52 | 53 | precision = float(TP) / float(TP + FP) 54 | recall = float(TP) / float(TP + FN) 55 | f1 = 2*precision*recall / (precision + recall) 56 | acc = (TP + TN) / (TP + TN + FP + FN) 57 | print('Accuracy: {}'.format(acc)) 58 | print('Precision: {}'.format(precision)) 59 | print('Recall: {}'.format(recall)) 60 | print('F1 score: {}'.format(f1)) 61 | print('Yes ratio: {}'.format(yes_ratio)) 62 | print('%.3f, %.3f, %.3f, %.3f, %.3f' % (f1, acc, precision, recall, yes_ratio) ) 63 | 64 | if __name__ == "__main__": 65 | parser = argparse.ArgumentParser() 66 | parser.add_argument("--annotation-dir", type=str) 67 | parser.add_argument("--question-file", type=str) 68 | parser.add_argument("--result-file", type=str) 69 | args = parser.parse_args() 70 | 71 | questions = [json.loads(line) for line in open(args.question_file)] 72 | questions = {question['question_id']: question for question in questions} 73 | answers = [json.loads(q) for q in open(args.result_file)] 74 | for file in os.listdir(args.annotation_dir): 75 | assert file.startswith('coco_pope_') 76 | assert file.endswith('.json') 77 | category = file[10:-5] 78 | cur_answers = [x for x in answers if questions[x['question_id']]['category'] == category] 79 | print('Category: {}, # samples: {}'.format(category, len(cur_answers))) 80 | eval_pope(cur_answers, os.path.join(args.annotation_dir, file)) 81 | print("====================================") 82 | -------------------------------------------------------------------------------- /llava/eval/eval_textvqa.py: -------------------------------------------------------------------------------- 1 | import os 2 | import argparse 3 | import json 4 | import re 5 | 6 | from llava.eval.m4c_evaluator import TextVQAAccuracyEvaluator 7 | 8 | 9 | def get_args(): 10 | parser = argparse.ArgumentParser() 11 | parser.add_argument('--annotation-file', type=str) 12 | parser.add_argument('--result-file', type=str) 13 | parser.add_argument('--result-dir', type=str) 14 | return parser.parse_args() 15 | 16 | 17 | def prompt_processor(prompt): 18 | if prompt.startswith('OCR tokens: '): 19 | pattern = r"Question: (.*?) Short answer:" 20 | match = re.search(pattern, prompt, re.DOTALL) 21 | question = match.group(1) 22 | elif 'Reference OCR token: ' in prompt and len(prompt.split('\n')) == 3: 23 | if prompt.startswith('Reference OCR token:'): 24 | question = prompt.split('\n')[1] 25 | else: 26 | question = prompt.split('\n')[0] 27 | elif len(prompt.split('\n')) == 2: 28 | question = prompt.split('\n')[0] 29 | else: 30 | assert False 31 | 32 | return question.lower() 33 | 34 | 35 | def eval_single(annotation_file, result_file): 36 | experiment_name = os.path.splitext(os.path.basename(result_file))[0] 37 | print(experiment_name) 38 | annotations = json.load(open(annotation_file))['data'] 39 | annotations = {(annotation['image_id'], annotation['question'].lower()): annotation for annotation in annotations} 40 | results = [json.loads(line) for line in open(result_file)] 41 | 42 | pred_list = [] 43 | for result in results: 44 | annotation = annotations[(result['question_id'], prompt_processor(result['prompt']))] 45 | pred_list.append({ 46 | "pred_answer": result['text'], 47 | "gt_answers": annotation['answers'], 48 | }) 49 | 50 | evaluator = TextVQAAccuracyEvaluator() 51 | print('Samples: {}\nAccuracy: {:.2f}%\n'.format(len(pred_list), 100. * evaluator.eval_pred_list(pred_list))) 52 | 53 | 54 | if __name__ == "__main__": 55 | args = get_args() 56 | 57 | if args.result_file is not None: 58 | eval_single(args.annotation_file, args.result_file) 59 | 60 | if args.result_dir is not None: 61 | for result_file in sorted(os.listdir(args.result_dir)): 62 | if not result_file.endswith('.jsonl'): 63 | print(f'Skipping {result_file}') 64 | continue 65 | eval_single(args.annotation_file, os.path.join(args.result_dir, result_file)) 66 | -------------------------------------------------------------------------------- /llava/eval/qa_baseline_gpt35.py: -------------------------------------------------------------------------------- 1 | """Generate answers with GPT-3.5""" 2 | # Note: you need to be using OpenAI Python v0.27.0 for the code below to work 3 | import argparse 4 | import json 5 | import os 6 | import time 7 | import concurrent.futures 8 | 9 | import openai 10 | import tqdm 11 | import shortuuid 12 | 13 | MODEL = 'gpt-3.5-turbo' 14 | MODEL_ID = 'gpt-3.5-turbo:20230327' 15 | 16 | def get_answer(question_id: int, question: str, max_tokens: int): 17 | ans = { 18 | 'answer_id': shortuuid.uuid(), 19 | 'question_id': question_id, 20 | 'model_id': MODEL_ID, 21 | } 22 | for _ in range(3): 23 | try: 24 | response = openai.ChatCompletion.create( 25 | model=MODEL, 26 | messages=[{ 27 | 'role': 'system', 28 | 'content': 'You are a helpful assistant.' 29 | }, { 30 | 'role': 'user', 31 | 'content': question, 32 | }], 33 | max_tokens=max_tokens, 34 | ) 35 | ans['text'] = response['choices'][0]['message']['content'] 36 | return ans 37 | except Exception as e: 38 | print('[ERROR]', e) 39 | ans['text'] = '#ERROR#' 40 | time.sleep(1) 41 | return ans 42 | 43 | 44 | if __name__ == '__main__': 45 | parser = argparse.ArgumentParser(description='ChatGPT answer generation.') 46 | parser.add_argument('-q', '--question') 47 | parser.add_argument('-o', '--output') 48 | parser.add_argument('--max-tokens', type=int, default=1024, help='maximum number of tokens produced in the output') 49 | args = parser.parse_args() 50 | 51 | questions_dict = {} 52 | with open(os.path.expanduser(args.question)) as f: 53 | for line in f: 54 | if not line: 55 | continue 56 | q = json.loads(line) 57 | questions_dict[q['question_id']] = q['text'] 58 | 59 | answers = [] 60 | 61 | with concurrent.futures.ThreadPoolExecutor(max_workers=32) as executor: 62 | futures = [] 63 | for qid, question in questions_dict.items(): 64 | future = executor.submit(get_answer, qid, question, args.max_tokens) 65 | futures.append(future) 66 | 67 | for future in tqdm.tqdm(concurrent.futures.as_completed(futures), total=len(futures)): 68 | answers.append(future.result()) 69 | 70 | answers.sort(key=lambda x: x['question_id']) 71 | 72 | with open(os.path.expanduser(args.output), 'w') as f: 73 | table = [json.dumps(ans) for ans in answers] 74 | f.write('\n'.join(table)) 75 | -------------------------------------------------------------------------------- /llava/eval/summarize_gpt_review.py: -------------------------------------------------------------------------------- 1 | import json 2 | import os 3 | from collections import defaultdict 4 | 5 | import numpy as np 6 | 7 | import argparse 8 | 9 | def parse_args(): 10 | parser = argparse.ArgumentParser(description='ChatGPT-based QA evaluation.') 11 | parser.add_argument('-d', '--dir', default=None) 12 | parser.add_argument('-v', '--version', default=None) 13 | parser.add_argument('-s', '--select', nargs='*', default=None) 14 | parser.add_argument('-f', '--files', nargs='*', default=[]) 15 | parser.add_argument('-i', '--ignore', nargs='*', default=[]) 16 | return parser.parse_args() 17 | 18 | 19 | if __name__ == '__main__': 20 | args = parse_args() 21 | 22 | if args.ignore is not None: 23 | args.ignore = [int(x) for x in args.ignore] 24 | 25 | if len(args.files) > 0: 26 | review_files = args.files 27 | else: 28 | review_files = [x for x in os.listdir(args.dir) if x.endswith('.jsonl') and (x.startswith('gpt4_text') or x.startswith('reviews_') or x.startswith('review_') or 'review' in args.dir)] 29 | 30 | for review_file in sorted(review_files): 31 | config = os.path.basename(review_file).replace('gpt4_text_', '').replace('.jsonl', '') 32 | if args.select is not None and any(x not in config for x in args.select): 33 | continue 34 | if '0613' in config: 35 | version = '0613' 36 | else: 37 | version = '0314' 38 | if args.version is not None and args.version != version: 39 | continue 40 | scores = defaultdict(list) 41 | print(config) 42 | with open(os.path.join(args.dir, review_file) if args.dir is not None else review_file) as f: 43 | for review_str in f: 44 | review = json.loads(review_str) 45 | if review['question_id'] in args.ignore: 46 | continue 47 | if 'category' in review: 48 | scores[review['category']].append(review['tuple']) 49 | scores['all'].append(review['tuple']) 50 | else: 51 | if 'tuple' in review: 52 | scores['all'].append(review['tuple']) 53 | else: 54 | scores['all'].append(review['score']) 55 | for k, v in sorted(scores.items()): 56 | stats = np.asarray(v).mean(0).tolist() 57 | stats = [round(x, 3) for x in stats] 58 | # print(k, stats, round(stats[1]/stats[0]*100, 1)) 59 | print(k, round(stats[1]/stats[0]*100, 1), round(stats[0] * 10, 1), round(stats[1] * 10, 1)) 60 | print('=================================') 61 | -------------------------------------------------------------------------------- /llava/eval/webpage/figures/alpaca.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/llava/eval/webpage/figures/alpaca.png -------------------------------------------------------------------------------- /llava/eval/webpage/figures/bard.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/llava/eval/webpage/figures/bard.jpg -------------------------------------------------------------------------------- /llava/eval/webpage/figures/chatgpt.svg: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /llava/eval/webpage/figures/llama.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/llava/eval/webpage/figures/llama.jpg -------------------------------------------------------------------------------- /llava/eval/webpage/figures/swords_FILL0_wght300_GRAD0_opsz48.svg: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /llava/eval/webpage/figures/vicuna.jpeg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/llava/eval/webpage/figures/vicuna.jpeg -------------------------------------------------------------------------------- /llava/eval/webpage/styles.css: -------------------------------------------------------------------------------- 1 | body { 2 | font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif; 3 | background-color: #f8f9fa; 4 | } 5 | 6 | .navbar-dark .navbar-nav .nav-link { 7 | color: #f1cf68; 8 | font-size: 1.1rem; 9 | padding: 0.5rem 0.6rem; 10 | } 11 | 12 | .card-header { 13 | font-weight: bold; 14 | } 15 | 16 | .card { 17 | box-shadow: 0 4px 8px rgba(0, 0, 0, 0.1); 18 | transition: 0.3s; 19 | } 20 | 21 | .card:hover { 22 | box-shadow: 0 8px 16px rgba(0, 0, 0, 0.2); 23 | } 24 | 25 | button { 26 | transition: background-color 0.3s; 27 | } 28 | 29 | button:hover { 30 | background-color: #007bff; 31 | } 32 | 33 | @media (max-width: 767px) { 34 | .form-row .form-group { 35 | margin-bottom: 10px; 36 | } 37 | } 38 | 39 | /* Extra styles */ 40 | 41 | .expandable-card .card-text-container { 42 | max-height: 200px; 43 | overflow-y: hidden; 44 | position: relative; 45 | } 46 | 47 | .expandable-card.expanded .card-text-container { 48 | max-height: none; 49 | } 50 | 51 | .expand-btn { 52 | position: relative; 53 | display: none; 54 | background-color: rgba(255, 255, 255, 0.8); 55 | color: #510c75; 56 | border-color: transparent; 57 | } 58 | 59 | .expand-btn:hover { 60 | background-color: rgba(200, 200, 200, 0.8); 61 | text-decoration: none; 62 | border-color: transparent; 63 | color: #510c75; 64 | } 65 | 66 | .expand-btn:focus { 67 | outline: none; 68 | text-decoration: none; 69 | } 70 | 71 | .expandable-card:not(.expanded) .card-text-container:after { 72 | content: ""; 73 | position: absolute; 74 | bottom: 0; 75 | left: 0; 76 | width: 100%; 77 | height: 90px; 78 | background: linear-gradient(rgba(255, 255, 255, 0.2), rgba(255, 255, 255, 1)); 79 | } 80 | 81 | .expandable-card:not(.expanded) .expand-btn { 82 | margin-top: -40px; 83 | } 84 | 85 | .card-body { 86 | padding-bottom: 5px; 87 | } 88 | 89 | .vertical-flex-layout { 90 | justify-content: center; 91 | align-items: center; 92 | height: 100%; 93 | display: flex; 94 | flex-direction: column; 95 | gap: 5px; 96 | } 97 | 98 | .figure-img { 99 | max-width: 100%; 100 | height: auto; 101 | } 102 | 103 | .adjustable-font-size { 104 | font-size: calc(0.5rem + 2vw); 105 | } 106 | -------------------------------------------------------------------------------- /llava/model/__init__.py: -------------------------------------------------------------------------------- 1 | from .language_model.llava_llama import LlavaLlamaForCausalLM, LlavaLlamaBPOForCausalLM, LlavaConfig, LlavaLlamaForSequenceClassification, LlavaLlamaForSequenceClassificationSep #, LlavaLlamaForCausalLMAdapt 2 | # from .language_model.llava_mpt import LlavaMPTForCausalLM, LlavaMPTConfig 3 | -------------------------------------------------------------------------------- /llava/model/apply_delta.py: -------------------------------------------------------------------------------- 1 | """ 2 | Usage: 3 | python3 -m fastchat.model.apply_delta --base ~/model_weights/llama-7b --target ~/model_weights/vicuna-7b --delta lmsys/vicuna-7b-delta 4 | """ 5 | import argparse 6 | 7 | import torch 8 | from tqdm import tqdm 9 | from transformers import AutoTokenizer, AutoModelForCausalLM 10 | from llava import LlavaLlamaForCausalLM 11 | 12 | 13 | def apply_delta(base_model_path, target_model_path, delta_path): 14 | print("Loading base model") 15 | base = AutoModelForCausalLM.from_pretrained( 16 | base_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True) 17 | 18 | print("Loading delta") 19 | delta = LlavaLlamaForCausalLM.from_pretrained(delta_path, torch_dtype=torch.float16, low_cpu_mem_usage=True) 20 | delta_tokenizer = AutoTokenizer.from_pretrained(delta_path) 21 | 22 | print("Applying delta") 23 | for name, param in tqdm(delta.state_dict().items(), desc="Applying delta"): 24 | if name not in base.state_dict(): 25 | assert name in ['model.mm_projector.weight', 'model.mm_projector.bias'], f'{name} not in base model' 26 | continue 27 | if param.data.shape == base.state_dict()[name].shape: 28 | param.data += base.state_dict()[name] 29 | else: 30 | assert name in ['model.embed_tokens.weight', 'lm_head.weight'], \ 31 | f'{name} dimension mismatch: {param.data.shape} vs {base.state_dict()[name].shape}' 32 | bparam = base.state_dict()[name] 33 | param.data[:bparam.shape[0], :bparam.shape[1]] += bparam 34 | 35 | print("Saving target model") 36 | delta.save_pretrained(target_model_path) 37 | delta_tokenizer.save_pretrained(target_model_path) 38 | 39 | 40 | if __name__ == "__main__": 41 | parser = argparse.ArgumentParser() 42 | parser.add_argument("--base-model-path", type=str, required=True) 43 | parser.add_argument("--target-model-path", type=str, required=True) 44 | parser.add_argument("--delta-path", type=str, required=True) 45 | 46 | args = parser.parse_args() 47 | 48 | apply_delta(args.base_model_path, args.target_model_path, args.delta_path) 49 | -------------------------------------------------------------------------------- /llava/model/consolidate.py: -------------------------------------------------------------------------------- 1 | """ 2 | Usage: 3 | python3 -m llava.model.consolidate --src ~/model_weights/llava-7b --dst ~/model_weights/llava-7b_consolidate 4 | """ 5 | import argparse 6 | 7 | import torch 8 | from transformers import AutoTokenizer, AutoModelForCausalLM 9 | from llava.model import * 10 | from llava.model.utils import auto_upgrade 11 | 12 | 13 | def consolidate_ckpt(src_path, dst_path): 14 | print("Loading model") 15 | auto_upgrade(src_path) 16 | src_model = AutoModelForCausalLM.from_pretrained(src_path, torch_dtype=torch.float16, low_cpu_mem_usage=True) 17 | src_tokenizer = AutoTokenizer.from_pretrained(src_path, use_fast=False) 18 | src_model.save_pretrained(dst_path) 19 | src_tokenizer.save_pretrained(dst_path) 20 | 21 | 22 | if __name__ == "__main__": 23 | parser = argparse.ArgumentParser() 24 | parser.add_argument("--src", type=str, required=True) 25 | parser.add_argument("--dst", type=str, required=True) 26 | 27 | args = parser.parse_args() 28 | 29 | consolidate_ckpt(args.src, args.dst) 30 | -------------------------------------------------------------------------------- /llava/model/language_model/mpt/adapt_tokenizer.py: -------------------------------------------------------------------------------- 1 | from typing import Union 2 | from transformers import AutoTokenizer, PreTrainedTokenizer, PreTrainedTokenizerFast 3 | Tokenizer = Union[PreTrainedTokenizer, PreTrainedTokenizerFast] 4 | NUM_SENTINEL_TOKENS: int = 100 5 | 6 | def adapt_tokenizer_for_denoising(tokenizer: Tokenizer): 7 | """Adds sentinel tokens and padding token (if missing). 8 | 9 | Expands the tokenizer vocabulary to include sentinel tokens 10 | used in mixture-of-denoiser tasks as well as a padding token. 11 | 12 | All added tokens are added as special tokens. No tokens are 13 | added if sentinel tokens and padding token already exist. 14 | """ 15 | sentinels_to_add = [f'' for i in range(NUM_SENTINEL_TOKENS)] 16 | tokenizer.add_tokens(sentinels_to_add, special_tokens=True) 17 | if tokenizer.pad_token is None: 18 | tokenizer.add_tokens('', special_tokens=True) 19 | tokenizer.pad_token = '' 20 | assert tokenizer.pad_token_id is not None 21 | sentinels = ''.join([f'' for i in range(NUM_SENTINEL_TOKENS)]) 22 | _sentinel_token_ids = tokenizer(sentinels, add_special_tokens=False).input_ids 23 | tokenizer.sentinel_token_ids = _sentinel_token_ids 24 | 25 | class AutoTokenizerForMOD(AutoTokenizer): 26 | """AutoTokenizer + Adaptation for MOD. 27 | 28 | A simple wrapper around AutoTokenizer to make instantiating 29 | an MOD-adapted tokenizer a bit easier. 30 | 31 | MOD-adapted tokenizers have sentinel tokens (e.g., ), 32 | a padding token, and a property to get the token ids of the 33 | sentinel tokens. 34 | """ 35 | 36 | @classmethod 37 | def from_pretrained(cls, *args, **kwargs): 38 | """See `AutoTokenizer.from_pretrained` docstring.""" 39 | tokenizer = super().from_pretrained(*args, **kwargs) 40 | adapt_tokenizer_for_denoising(tokenizer) 41 | return tokenizer -------------------------------------------------------------------------------- /llava/model/language_model/mpt/blocks.py: -------------------------------------------------------------------------------- 1 | """GPT Blocks used for the GPT Model.""" 2 | from typing import Dict, Optional, Tuple 3 | import torch 4 | import torch.nn as nn 5 | from .attention import ATTN_CLASS_REGISTRY 6 | from .norm import NORM_CLASS_REGISTRY 7 | 8 | class MPTMLP(nn.Module): 9 | 10 | def __init__(self, d_model: int, expansion_ratio: int, device: Optional[str]=None): 11 | super().__init__() 12 | self.up_proj = nn.Linear(d_model, expansion_ratio * d_model, device=device) 13 | self.act = nn.GELU(approximate='none') 14 | self.down_proj = nn.Linear(expansion_ratio * d_model, d_model, device=device) 15 | self.down_proj._is_residual = True 16 | 17 | def forward(self, x): 18 | return self.down_proj(self.act(self.up_proj(x))) 19 | 20 | class MPTBlock(nn.Module): 21 | 22 | def __init__(self, d_model: int, n_heads: int, expansion_ratio: int, attn_config: Dict={'attn_type': 'multihead_attention', 'attn_pdrop': 0.0, 'attn_impl': 'triton', 'qk_ln': False, 'clip_qkv': None, 'softmax_scale': None, 'prefix_lm': False, 'attn_uses_sequence_id': False, 'alibi': False, 'alibi_bias_max': 8}, resid_pdrop: float=0.0, norm_type: str='low_precision_layernorm', verbose: int=0, device: Optional[str]=None, **kwargs): 23 | del kwargs 24 | super().__init__() 25 | norm_class = NORM_CLASS_REGISTRY[norm_type.lower()] 26 | attn_class = ATTN_CLASS_REGISTRY[attn_config['attn_type']] 27 | self.norm_1 = norm_class(d_model, device=device) 28 | self.attn = attn_class(attn_impl=attn_config['attn_impl'], clip_qkv=attn_config['clip_qkv'], qk_ln=attn_config['qk_ln'], softmax_scale=attn_config['softmax_scale'], attn_pdrop=attn_config['attn_pdrop'], d_model=d_model, n_heads=n_heads, verbose=verbose, device=device) 29 | self.norm_2 = norm_class(d_model, device=device) 30 | self.ffn = MPTMLP(d_model=d_model, expansion_ratio=expansion_ratio, device=device) 31 | self.resid_attn_dropout = nn.Dropout(resid_pdrop) 32 | self.resid_ffn_dropout = nn.Dropout(resid_pdrop) 33 | 34 | def forward(self, x: torch.Tensor, past_key_value: Optional[Tuple[torch.Tensor]]=None, attn_bias: Optional[torch.Tensor]=None, attention_mask: Optional[torch.ByteTensor]=None, is_causal: bool=True) -> Tuple[torch.Tensor, Optional[Tuple[torch.Tensor]]]: 35 | a = self.norm_1(x) 36 | (b, attn_weights, past_key_value) = self.attn(a, past_key_value=past_key_value, attn_bias=attn_bias, attention_mask=attention_mask, is_causal=is_causal) 37 | x = x + self.resid_attn_dropout(b) 38 | m = self.norm_2(x) 39 | n = self.ffn(m) 40 | x = x + self.resid_ffn_dropout(n) 41 | return (x, attn_weights, past_key_value) -------------------------------------------------------------------------------- /llava/model/language_model/mpt/custom_embedding.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn as nn 3 | import torch.nn.functional as F 4 | from torch import Tensor 5 | 6 | class SharedEmbedding(nn.Embedding): 7 | 8 | def forward(self, input: Tensor, unembed: bool=False) -> Tensor: 9 | if unembed: 10 | return F.linear(input, self.weight) 11 | return super().forward(input) -------------------------------------------------------------------------------- /llava/model/language_model/mpt/norm.py: -------------------------------------------------------------------------------- 1 | import torch 2 | 3 | def _cast_if_autocast_enabled(tensor): 4 | if torch.is_autocast_enabled(): 5 | if tensor.device.type == 'cuda': 6 | dtype = torch.get_autocast_gpu_dtype() 7 | elif tensor.device.type == 'cpu': 8 | dtype = torch.get_autocast_cpu_dtype() 9 | else: 10 | raise NotImplementedError() 11 | return tensor.to(dtype=dtype) 12 | return tensor 13 | 14 | class LPLayerNorm(torch.nn.LayerNorm): 15 | 16 | def __init__(self, normalized_shape, eps=1e-05, elementwise_affine=True, device=None, dtype=None): 17 | super().__init__(normalized_shape=normalized_shape, eps=eps, elementwise_affine=elementwise_affine, device=device, dtype=dtype) 18 | 19 | def forward(self, x): 20 | module_device = x.device 21 | downcast_x = _cast_if_autocast_enabled(x) 22 | downcast_weight = _cast_if_autocast_enabled(self.weight) if self.weight is not None else self.weight 23 | downcast_bias = _cast_if_autocast_enabled(self.bias) if self.bias is not None else self.bias 24 | with torch.autocast(enabled=False, device_type=module_device.type): 25 | return torch.nn.functional.layer_norm(downcast_x, self.normalized_shape, downcast_weight, downcast_bias, self.eps) 26 | 27 | def rms_norm(x, weight=None, eps=1e-05): 28 | output = x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + eps) 29 | if weight is not None: 30 | return output * weight 31 | return output 32 | 33 | class RMSNorm(torch.nn.Module): 34 | 35 | def __init__(self, normalized_shape, eps=1e-05, weight=True, dtype=None, device=None): 36 | super().__init__() 37 | self.eps = eps 38 | if weight: 39 | self.weight = torch.nn.Parameter(torch.ones(normalized_shape, dtype=dtype, device=device)) 40 | else: 41 | self.register_parameter('weight', None) 42 | 43 | def forward(self, x): 44 | return rms_norm(x.float(), self.weight, self.eps).to(dtype=x.dtype) 45 | 46 | class LPRMSNorm(RMSNorm): 47 | 48 | def __init__(self, normalized_shape, eps=1e-05, weight=True, dtype=None, device=None): 49 | super().__init__(normalized_shape=normalized_shape, eps=eps, weight=weight, dtype=dtype, device=device) 50 | 51 | def forward(self, x): 52 | downcast_x = _cast_if_autocast_enabled(x) 53 | downcast_weight = _cast_if_autocast_enabled(self.weight) if self.weight is not None else self.weight 54 | with torch.autocast(enabled=False, device_type=x.device.type): 55 | return rms_norm(downcast_x, downcast_weight, self.eps).to(dtype=x.dtype) 56 | NORM_CLASS_REGISTRY = {'layernorm': torch.nn.LayerNorm, 'low_precision_layernorm': LPLayerNorm, 'rmsnorm': RMSNorm, 'low_precision_rmsnorm': LPRMSNorm} -------------------------------------------------------------------------------- /llava/model/make_delta.py: -------------------------------------------------------------------------------- 1 | """ 2 | Usage: 3 | python3 -m llava.model.make_delta --base ~/model_weights/llama-7b --target ~/model_weights/llava-7b --delta ~/model_weights/llava-7b-delta --hub-repo-id liuhaotian/llava-7b-delta 4 | """ 5 | import argparse 6 | 7 | import torch 8 | from tqdm import tqdm 9 | from transformers import AutoTokenizer, AutoModelForCausalLM 10 | from llava.model.utils import auto_upgrade 11 | 12 | 13 | def make_delta(base_model_path, target_model_path, delta_path, hub_repo_id): 14 | print("Loading base model") 15 | base = AutoModelForCausalLM.from_pretrained( 16 | base_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True) 17 | 18 | print("Loading target model") 19 | auto_upgrade(target_model_path) 20 | target = AutoModelForCausalLM.from_pretrained(target_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True) 21 | 22 | print("Calculating delta") 23 | for name, param in tqdm(target.state_dict().items(), desc="Calculating delta"): 24 | if name not in base.state_dict(): 25 | assert name in ['model.mm_projector.weight', 'model.mm_projector.bias'], f'{name} not in base model' 26 | continue 27 | if param.data.shape == base.state_dict()[name].shape: 28 | param.data -= base.state_dict()[name] 29 | else: 30 | assert name in ['model.embed_tokens.weight', 'lm_head.weight'], f'{name} dimension mismatch: {param.data.shape} vs {base.state_dict()[name].shape}' 31 | bparam = base.state_dict()[name] 32 | param.data[:bparam.shape[0], :bparam.shape[1]] -= bparam 33 | 34 | print("Saving delta") 35 | if hub_repo_id: 36 | kwargs = {"push_to_hub": True, "repo_id": hub_repo_id} 37 | else: 38 | kwargs = {} 39 | target.save_pretrained(delta_path, **kwargs) 40 | target_tokenizer = AutoTokenizer.from_pretrained(target_model_path) 41 | target_tokenizer.save_pretrained(delta_path, **kwargs) 42 | 43 | 44 | if __name__ == "__main__": 45 | parser = argparse.ArgumentParser() 46 | parser.add_argument("--base-model-path", type=str, required=True) 47 | parser.add_argument("--target-model-path", type=str, required=True) 48 | parser.add_argument("--delta-path", type=str, required=True) 49 | parser.add_argument("--hub-repo-id", type=str, default=None) 50 | args = parser.parse_args() 51 | 52 | make_delta(args.base_model_path, args.target_model_path, args.delta_path, args.hub_repo_id) 53 | -------------------------------------------------------------------------------- /llava/model/multimodal_encoder/builder.py: -------------------------------------------------------------------------------- 1 | import os 2 | from .clip_encoder import CLIPVisionTower 3 | 4 | 5 | def build_vision_tower(vision_tower_cfg, **kwargs): 6 | vision_tower = getattr(vision_tower_cfg, 'mm_vision_tower', getattr(vision_tower_cfg, 'vision_tower', None)) 7 | is_absolute_path_exists = os.path.exists(vision_tower) 8 | if is_absolute_path_exists or vision_tower.startswith("openai") or vision_tower.startswith("laion"): 9 | return CLIPVisionTower(vision_tower, args=vision_tower_cfg, **kwargs) 10 | 11 | raise ValueError(f'Unknown vision tower: {vision_tower}') 12 | -------------------------------------------------------------------------------- /llava/model/multimodal_encoder/clip_encoder.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn as nn 3 | 4 | from transformers import CLIPVisionModel, CLIPImageProcessor, CLIPVisionConfig 5 | 6 | 7 | class CLIPVisionTower(nn.Module): 8 | def __init__(self, vision_tower, args, delay_load=False): 9 | super().__init__() 10 | 11 | self.is_loaded = False 12 | 13 | self.vision_tower_name = vision_tower 14 | self.select_layer = args.mm_vision_select_layer 15 | self.select_feature = getattr(args, 'mm_vision_select_feature', 'patch') 16 | 17 | if not delay_load: 18 | self.load_model() 19 | else: 20 | self.cfg_only = CLIPVisionConfig.from_pretrained(self.vision_tower_name) 21 | 22 | def load_model(self): 23 | self.image_processor = CLIPImageProcessor.from_pretrained(self.vision_tower_name) 24 | self.vision_tower = CLIPVisionModel.from_pretrained(self.vision_tower_name) 25 | self.vision_tower.requires_grad_(False) 26 | 27 | self.is_loaded = True 28 | 29 | def feature_select(self, image_forward_outs): 30 | image_features = image_forward_outs.hidden_states[self.select_layer] 31 | if self.select_feature == 'patch': 32 | image_features = image_features[:, 1:] 33 | elif self.select_feature == 'cls_patch': 34 | image_features = image_features 35 | else: 36 | raise ValueError(f'Unexpected select feature: {self.select_feature}') 37 | return image_features 38 | 39 | @torch.no_grad() 40 | def forward(self, images): 41 | if type(images) is list: 42 | image_features = [] 43 | for image in images: 44 | image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), output_hidden_states=True) 45 | image_feature = self.feature_select(image_forward_out).to(image.dtype) 46 | image_features.append(image_feature) 47 | else: 48 | image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), output_hidden_states=True) 49 | image_features = self.feature_select(image_forward_outs).to(images.dtype) 50 | 51 | return image_features 52 | 53 | @property 54 | def dummy_feature(self): 55 | return torch.zeros(1, self.hidden_size, device=self.device, dtype=self.dtype) 56 | 57 | @property 58 | def dtype(self): 59 | return self.vision_tower.dtype 60 | 61 | @property 62 | def device(self): 63 | return self.vision_tower.device 64 | 65 | @property 66 | def config(self): 67 | if self.is_loaded: 68 | return self.vision_tower.config 69 | else: 70 | return self.cfg_only 71 | 72 | @property 73 | def hidden_size(self): 74 | return self.config.hidden_size 75 | 76 | @property 77 | def num_patches(self): 78 | return (self.config.image_size // self.config.patch_size) ** 2 79 | -------------------------------------------------------------------------------- /llava/model/multimodal_projector/builder.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn as nn 3 | import re 4 | 5 | 6 | class IdentityMap(nn.Module): 7 | def __init__(self): 8 | super().__init__() 9 | 10 | def forward(self, x, *args, **kwargs): 11 | return x 12 | 13 | @property 14 | def config(self): 15 | return {"mm_projector_type": 'identity'} 16 | 17 | 18 | class SimpleResBlock(nn.Module): 19 | def __init__(self, channels): 20 | super().__init__() 21 | self.pre_norm = nn.LayerNorm(channels) 22 | 23 | self.proj = nn.Sequential( 24 | nn.Linear(channels, channels), 25 | nn.GELU(), 26 | nn.Linear(channels, channels) 27 | ) 28 | def forward(self, x): 29 | x = self.pre_norm(x) 30 | return x + self.proj(x) 31 | 32 | 33 | def build_vision_projector(config, delay_load=False, **kwargs): 34 | projector_type = getattr(config, 'mm_projector_type', 'linear') 35 | 36 | if projector_type == 'linear': 37 | return nn.Linear(config.mm_hidden_size, config.hidden_size) 38 | 39 | mlp_gelu_match = re.match(r'^mlp(\d+)x_gelu$', projector_type) 40 | if mlp_gelu_match: 41 | mlp_depth = int(mlp_gelu_match.group(1)) 42 | modules = [nn.Linear(config.mm_hidden_size, config.hidden_size)] 43 | for _ in range(1, mlp_depth): 44 | modules.append(nn.GELU()) 45 | modules.append(nn.Linear(config.hidden_size, config.hidden_size)) 46 | return nn.Sequential(*modules) 47 | 48 | if projector_type == 'identity': 49 | return IdentityMap() 50 | 51 | raise ValueError(f'Unknown projector type: {projector_type}') 52 | 53 | 54 | def build_vision_projector_adap(config, delay_load=False, **kwargs): 55 | projector_type = getattr(config, 'mm_projector_type', 'linear') 56 | 57 | if projector_type == 'linear': 58 | return nn.Linear(config.mm_hidden_size, config.hidden_size) 59 | 60 | mlp_gelu_match = re.match(r'^mlp(\d+)x_gelu$', projector_type) 61 | if mlp_gelu_match: 62 | mlp_depth = int(mlp_gelu_match.group(1)) 63 | modules = [nn.Linear(config.mm_hidden_size, config.hidden_size)] 64 | for _ in range(1, mlp_depth): 65 | modules.append(nn.GELU()) 66 | modules.append(nn.Linear(config.hidden_size, config.hidden_size)) 67 | return nn.Sequential(*modules) 68 | 69 | if projector_type == 'identity': 70 | return IdentityMap() 71 | 72 | raise ValueError(f'Unknown projector type: {projector_type}') 73 | -------------------------------------------------------------------------------- /llava/model/utils.py: -------------------------------------------------------------------------------- 1 | from transformers import AutoConfig 2 | 3 | 4 | def auto_upgrade(config): 5 | cfg = AutoConfig.from_pretrained(config) 6 | if 'llava' in config and 'llava' not in cfg.model_type: 7 | assert cfg.model_type == 'llama' 8 | print("You are using newer LLaVA code base, while the checkpoint of v0 is from older code base.") 9 | print("You must upgrade the checkpoint to the new code base (this can be done automatically).") 10 | confirm = input("Please confirm that you want to upgrade the checkpoint. [Y/N]") 11 | if confirm.lower() in ["y", "yes"]: 12 | print("Upgrading checkpoint...") 13 | assert len(cfg.architectures) == 1 14 | setattr(cfg.__class__, "model_type", "llava") 15 | cfg.architectures[0] = 'LlavaLlamaForCausalLM' 16 | cfg.save_pretrained(config) 17 | print("Checkpoint upgraded.") 18 | else: 19 | print("Checkpoint upgrade aborted.") 20 | exit(1) 21 | -------------------------------------------------------------------------------- /llava/serve/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/llava/serve/__init__.py -------------------------------------------------------------------------------- /llava/serve/examples/extreme_ironing.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/llava/serve/examples/extreme_ironing.jpg -------------------------------------------------------------------------------- /llava/serve/examples/waterview.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/llava/serve/examples/waterview.jpg -------------------------------------------------------------------------------- /llava/serve/register_worker.py: -------------------------------------------------------------------------------- 1 | """ 2 | Manually register workers. 3 | 4 | Usage: 5 | python3 -m fastchat.serve.register_worker --controller http://localhost:21001 --worker-name http://localhost:21002 6 | """ 7 | 8 | import argparse 9 | 10 | import requests 11 | 12 | if __name__ == "__main__": 13 | parser = argparse.ArgumentParser() 14 | parser.add_argument("--controller-address", type=str) 15 | parser.add_argument("--worker-name", type=str) 16 | parser.add_argument("--check-heart-beat", action="store_true") 17 | args = parser.parse_args() 18 | 19 | url = args.controller_address + "/register_worker" 20 | data = { 21 | "worker_name": args.worker_name, 22 | "check_heart_beat": args.check_heart_beat, 23 | "worker_status": None, 24 | } 25 | r = requests.post(url, json=data) 26 | assert r.status_code == 200 27 | -------------------------------------------------------------------------------- /llava/serve/test_message.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | import json 3 | 4 | import requests 5 | 6 | from llava.conversation import default_conversation 7 | 8 | 9 | def main(): 10 | if args.worker_address: 11 | worker_addr = args.worker_address 12 | else: 13 | controller_addr = args.controller_address 14 | ret = requests.post(controller_addr + "/refresh_all_workers") 15 | ret = requests.post(controller_addr + "/list_models") 16 | models = ret.json()["models"] 17 | models.sort() 18 | print(f"Models: {models}") 19 | 20 | ret = requests.post(controller_addr + "/get_worker_address", 21 | json={"model": args.model_name}) 22 | worker_addr = ret.json()["address"] 23 | print(f"worker_addr: {worker_addr}") 24 | 25 | if worker_addr == "": 26 | return 27 | 28 | conv = default_conversation.copy() 29 | conv.append_message(conv.roles[0], args.message) 30 | prompt = conv.get_prompt() 31 | 32 | headers = {"User-Agent": "LLaVA Client"} 33 | pload = { 34 | "model": args.model_name, 35 | "prompt": prompt, 36 | "max_new_tokens": args.max_new_tokens, 37 | "temperature": 0.7, 38 | "stop": conv.sep, 39 | } 40 | response = requests.post(worker_addr + "/worker_generate_stream", headers=headers, 41 | json=pload, stream=True) 42 | 43 | print(prompt.replace(conv.sep, "\n"), end="") 44 | for chunk in response.iter_lines(chunk_size=8192, decode_unicode=False, delimiter=b"\0"): 45 | if chunk: 46 | data = json.loads(chunk.decode("utf-8")) 47 | output = data["text"].split(conv.sep)[-1] 48 | print(output, end="\r") 49 | print("") 50 | 51 | 52 | if __name__ == "__main__": 53 | parser = argparse.ArgumentParser() 54 | parser.add_argument("--controller-address", type=str, default="http://localhost:21001") 55 | parser.add_argument("--worker-address", type=str) 56 | parser.add_argument("--model-name", type=str, default="facebook/opt-350m") 57 | parser.add_argument("--max-new-tokens", type=int, default=32) 58 | parser.add_argument("--message", type=str, default= 59 | "Tell me a story with more than 1000 words.") 60 | args = parser.parse_args() 61 | 62 | main() 63 | -------------------------------------------------------------------------------- /llava/train/train_mem.py: -------------------------------------------------------------------------------- 1 | # Adopted from https://github.com/lm-sys/FastChat. Below is the original copyright: 2 | # Adopted from tatsu-lab@stanford_alpaca. Below is the original copyright: 3 | # Make it more memory efficient by monkey patching the LLaMA model with FlashAttn. 4 | 5 | # Need to call this before importing transformers. 6 | from llava.train.llama_flash_attn_monkey_patch import replace_llama_attn_with_flash_attn 7 | 8 | replace_llama_attn_with_flash_attn() 9 | 10 | from llava.train.train import train 11 | 12 | if __name__ == "__main__": 13 | train() 14 | -------------------------------------------------------------------------------- /llava/train/train_xformers.py: -------------------------------------------------------------------------------- 1 | # Make it more memory efficient by monkey patching the LLaMA model with xformers attention. 2 | 3 | # Need to call this before importing transformers. 4 | from llava.train.llama_xformers_attn_monkey_patch import ( 5 | replace_llama_attn_with_xformers_attn, 6 | ) 7 | 8 | replace_llama_attn_with_xformers_attn() 9 | 10 | from llava.train.train import train 11 | 12 | if __name__ == "__main__": 13 | train() 14 | -------------------------------------------------------------------------------- /main.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/main.py -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_0.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_1.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_100.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_100.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_101.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_101.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_102.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_102.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_103.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_103.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_104.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_104.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_105.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_105.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_106.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_106.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_107.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_107.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_108.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_108.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_109.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_109.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_11.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_11.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_110.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_110.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_111.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_111.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_112.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_112.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_113.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_113.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_114.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_114.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_115.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_115.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_116.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_116.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_117.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_117.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_118.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_118.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_119.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_119.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_120.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_120.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_121.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_121.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_122.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_122.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_123.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_123.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_124.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_124.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_125.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_125.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_126.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_126.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_127.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_127.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_128.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_128.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_129.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_129.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_13.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_13.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_130.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_130.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_131.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_131.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_132.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_132.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_133.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_133.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_134.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_134.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_135.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_135.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_136.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_136.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_137.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_137.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_138.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_138.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_139.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_139.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_140.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_140.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_141.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_141.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_142.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_142.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_143.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_143.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_144.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_144.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_145.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_145.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_146.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_146.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_147.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_147.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_148.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_148.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_149.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_149.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_15.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_150.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_150.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_151.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_151.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_152.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_152.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_153.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_153.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_154.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_154.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_155.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_155.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_156.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_156.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_157.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_157.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_158.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_158.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_159.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_159.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_16.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_16.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_160.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_160.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_161.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_161.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_162.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_162.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_163.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_163.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_164.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_164.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_165.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_165.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_166.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_166.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_167.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_167.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_168.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_168.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_169.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_169.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_17.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_17.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_170.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_170.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_171.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_171.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_172.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_172.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_173.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_173.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_174.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_174.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_175.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_175.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_176.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_176.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_177.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_177.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_178.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_178.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_179.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_179.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_18.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_18.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_180.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_180.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_181.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_181.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_182.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_182.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_183.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_183.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_184.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_184.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_185.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_185.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_186.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_186.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_187.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_187.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_188.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_188.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_189.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_189.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_190.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_190.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_191.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_191.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_192.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_192.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_193.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_193.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_194.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_194.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_195.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_195.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_196.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_196.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_197.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_197.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_198.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_198.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_199.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_199.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_20.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_20.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_200.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_200.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_201.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_201.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_202.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_202.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_203.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_203.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_204.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_204.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_205.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_205.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_206.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_206.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_207.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_207.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_208.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_208.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_209.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_209.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_21.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_21.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_210.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_210.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_211.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_211.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_212.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_212.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_213.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_213.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_214.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_214.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_215.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_215.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_216.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_216.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_217.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_217.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_23.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_23.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_25.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_25.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_27.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_27.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_28.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_28.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_3.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_30.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_30.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_31.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_31.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_32.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_32.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_33.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_33.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_34.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_34.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_35.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_35.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_36.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_36.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_37.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_37.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_38.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_38.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_39.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_39.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_40.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_40.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_41.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_41.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_42.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_42.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_44.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_44.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_46.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_46.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_48.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_48.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_49.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_49.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_5.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_50.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_50.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_52.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_52.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_53.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_53.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_54.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_54.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_55.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_55.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_56.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_56.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_57.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_57.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_58.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_58.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_60.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_60.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_62.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_62.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_64.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_64.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_65.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_65.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_66.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_66.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_67.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_67.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_68.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_68.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_69.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_69.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_7.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_70.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_70.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_71.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_71.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_72.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_72.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_73.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_73.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_74.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_74.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_75.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_75.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_76.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_76.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_77.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_77.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_78.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_78.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_79.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_79.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_8.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_80.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_80.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_81.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_81.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_82.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_82.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_83.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_83.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_84.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_84.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_85.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_85.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_86.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_86.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_87.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_87.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_88.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_88.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_89.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_89.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_9.png -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_90.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_90.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_91.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_91.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_92.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_92.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_93.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_93.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_94.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_94.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_95.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_95.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_96.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_96.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_97.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_97.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_98.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_98.jpg -------------------------------------------------------------------------------- /playground/data/eval/mmvet_images/v1_99.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/pipilurj/bootstrapped-preference-optimization-BPO/ecaf28067e2ddb7ae3b53b7c1dd63fbfff33e89c/playground/data/eval/mmvet_images/v1_99.png -------------------------------------------------------------------------------- /pyproject.toml: -------------------------------------------------------------------------------- 1 | [build-system] 2 | requires = ["setuptools>=61.0"] 3 | build-backend = "setuptools.build_meta" 4 | 5 | [project] 6 | name = "llava" 7 | version = "1.1.3" 8 | description = "Towards GPT-4 like large language and visual assistant." 9 | readme = "README.md" 10 | requires-python = ">=3.8" 11 | classifiers = [ 12 | "Programming Language :: Python :: 3", 13 | "License :: OSI Approved :: Apache Software License", 14 | ] 15 | dependencies = [ 16 | "torch==2.1.2", "torchvision==0.16.2", 17 | "transformers==4.31.0", "tokenizers>=0.12.1,<0.14", "sentencepiece==0.1.99", "shortuuid", 18 | "accelerate==0.21.0", "peft==0.4.0", "bitsandbytes==0.41.0", 19 | "pydantic<2,>=1", "markdown2[all]", "numpy", "scikit-learn==1.2.2", 20 | "gradio==3.35.2", "gradio_client==0.2.9", 21 | "requests", "httpx==0.24.0", "uvicorn", "fastapi", 22 | "einops==0.6.1", "einops-exts==0.0.4", "timm==0.6.13", 23 | "trl==0.7.2" 24 | ] 25 | 26 | [project.optional-dependencies] 27 | train = ["deepspeed==0.9.5", "ninja", "wandb"] 28 | 29 | [project.urls] 30 | "Homepage" = "https://llava-vl.github.io" 31 | "Bug Tracker" = "https://github.com/haotian-liu/LLaVA/issues" 32 | 33 | [tool.setuptools.packages.find] 34 | exclude = ["assets*", "benchmark*", "docs", "dist*", "playground*", "scripts*", "tests*"] 35 | 36 | [tool.wheel] 37 | exclude = ["assets*", "benchmark*", "docs", "dist*", "playground*", "scripts*", "tests*"] 38 | -------------------------------------------------------------------------------- /qwen/README.md: -------------------------------------------------------------------------------- 1 | # BPO for Qwen 2 | 3 | Modified code from [[VLFeedback]](https://github.com/vlf-silkie/VLFeedback). Perform BPO on Qwen-VL-Chat using BPO data. 4 | 5 | ### Training data 6 | Download ShareGPT4V from [here](https://huggingface.co/datasets/Lin-Chen/ShareGPT4V) 7 | 8 | Download COCO from [here](https://cocodataset.org/#home) 9 | 10 | Download dataset annotation from [here](https://huggingface.co/datasets/renjiepi/BPO) 11 | 12 | Extract data from ShareGPT4V and organize the images as follows: 13 | 14 | ``` 15 | Image_root 16 | ├── coco/ 17 | train2017/ 18 | ├── llava/ 19 | llava_pretrain / 20 | ├── sam/ 21 | ├── share_textvqa/ 22 | images/ 23 | ├── web-celebrity/ 24 | images/ 25 | ├── web-landmark/ 26 | images/ 27 | ├── wikiart/ 28 | images/ 29 | ``` 30 | 31 | ### Installation 32 | 33 | To run our training scripts, create a virtual environment and install the dependencies first. 34 | 35 | ```bash 36 | conda create -n silkie python=3.10 && conda activate silkie 37 | pip install -r requirements.txt 38 | ``` 39 | 40 | ### Training 41 | 42 | Our training scripts support both single-node and multi-node training. 43 | We provide a `launch_dpo.py` script that handles both cases. If you want to launch a job locally, you can use: 44 | 45 | ```bash 46 | python launch_dpo.py --config dpo_config/example.yaml --working $WORKING_DIR 47 | ``` 48 | 49 | If you want to launch a job on a Slurm cluster, specify `GPUS_PER_NODE` in `launch_dpo.py` and run: 50 | 51 | ```bash 52 | python launch_dpo.py --config dpo_config/example.yaml --working $WORKING_DIR --gpus $NUM_GPUS 53 | ``` 54 | 55 | ## Citations 56 | 57 | ```bib 58 | @article{2023vlfeedback, 59 | author = {Lei Li and Zhihui Xie and Mukai Li and Shunian Chen and Peiyi Wang and Liang Chen and Yazheng Yang and Benyou Wang and Lingpeng Kong}, 60 | title = {Silkie: Preference Distillation for Large Visual Language Models}, 61 | publisher = {arXiv:2312.10665}, 62 | year = {2023} 63 | } 64 | ``` 65 | 66 | ## Acknowledgements 67 | 68 | We would like to thank the authors of [trl](https://github.com/huggingface/trl) and [Qwen-VL](https://github.com/QwenLM/Qwen-VL) for their great work. 69 | -------------------------------------------------------------------------------- /qwen/dpo_config/example.yaml: -------------------------------------------------------------------------------- 1 | model_name_or_path: "Qwen/Qwen-VL-Chat" 2 | output_dir: null # to be set by the script 3 | bf16: true 4 | fix_vit: true 5 | num_train_epochs: 3 6 | per_device_train_batch_size: 2 7 | per_device_eval_batch_size: 2 8 | gradient_accumulation_steps: 8 9 | evaluation_strategy: "steps" 10 | eval_steps: 500 11 | save_strategy: "steps" 12 | save_steps: 100 13 | save_total_limit: 10 14 | learning_rate: 1e-5 15 | weight_decay: 0.05 16 | adam_beta2: 0.98 17 | warmup_ratio: 0.1 18 | lr_scheduler_type: "cosine" 19 | logging_steps: 10 20 | report_to: wandb 21 | run_name: silkie-paperconfig 22 | model_max_length: 2048 23 | gradient_checkpointing: true 24 | use_lora: true 25 | bf16: true 26 | tf32: true 27 | logging_first_step: true 28 | remove_unused_columns: false 29 | -------------------------------------------------------------------------------- /qwen/requirements.txt: -------------------------------------------------------------------------------- 1 | accelerate==0.23.0 2 | datasets==2.14.6 3 | deepspeed==0.11.0 4 | numpy==1.26.2 5 | peft==0.5.0 6 | PyYAML==6.0.1 7 | submitit==1.5.1 8 | torch==2.0.1 9 | torchvision==0.15.2 10 | transformers==4.32.1 11 | trl==0.7.2 12 | einops 13 | tiktoken 14 | matplotlib 15 | pillow 16 | transformers_stream_generator 17 | wandb 18 | -------------------------------------------------------------------------------- /scripts/concatenate_json.py: -------------------------------------------------------------------------------- 1 | import sys 2 | import json 3 | 4 | def merge_json(files): 5 | merged_data = {} 6 | for file in files: 7 | with open(file, 'r') as f: 8 | data = json.load(f) 9 | merged_data.update(data) 10 | return merged_data 11 | 12 | if __name__ == "__main__": 13 | files = sys.argv[1:] 14 | merged_data = merge_json(files) 15 | with open('merge.json', 'w') as f: 16 | json.dump(merged_data, f, indent=4) 17 | -------------------------------------------------------------------------------- /scripts/convert_gqa_for_eval.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | import argparse 4 | 5 | parser = argparse.ArgumentParser() 6 | parser.add_argument("--src", type=str) 7 | parser.add_argument("--dst", type=str) 8 | args = parser.parse_args() 9 | 10 | all_answers = [] 11 | for line_idx, line in enumerate(open(args.src)): 12 | res = json.loads(line) 13 | question_id = res['question_id'] 14 | text = res['text'].rstrip('.').lower() 15 | all_answers.append({"questionId": question_id, "prediction": text}) 16 | 17 | with open(args.dst, 'w') as f: 18 | json.dump(all_answers, f) 19 | -------------------------------------------------------------------------------- /scripts/convert_mmbench_for_submission.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | import argparse 4 | import pandas as pd 5 | 6 | def get_args(): 7 | parser = argparse.ArgumentParser() 8 | parser.add_argument("--annotation-file", type=str, required=True) 9 | parser.add_argument("--result-dir", type=str, required=True) 10 | parser.add_argument("--upload-dir", type=str, required=True) 11 | parser.add_argument("--experiment", type=str, required=True) 12 | 13 | return parser.parse_args() 14 | 15 | if __name__ == "__main__": 16 | args = get_args() 17 | 18 | df = pd.read_table(args.annotation_file) 19 | 20 | cur_df = df.copy() 21 | cur_df = cur_df.drop(columns=['hint', 'category', 'source', 'image', 'comment', 'l2-category']) 22 | cur_df.insert(6, 'prediction', None) 23 | for pred in open(os.path.join(args.result_dir, f"{args.experiment}.jsonl")): 24 | pred = json.loads(pred) 25 | cur_df.loc[df['index'] == pred['question_id'], 'prediction'] = pred['text'] 26 | 27 | cur_df.to_excel(os.path.join(args.upload_dir, f"{args.experiment}.xlsx"), index=False, engine='openpyxl') 28 | -------------------------------------------------------------------------------- /scripts/convert_mmvet_for_eval.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | import argparse 4 | 5 | parser = argparse.ArgumentParser() 6 | parser.add_argument("--src", type=str) 7 | parser.add_argument("--dst", type=str) 8 | args = parser.parse_args() 9 | 10 | cur_result = {} 11 | 12 | for line in open(args.src): 13 | data = json.loads(line) 14 | qid = data['question_id'] 15 | # cur_result[f'v1_{qid}'] = data['text'] 16 | cur_result[qid] = data['text'] 17 | 18 | with open(args.dst, 'w') as f: 19 | json.dump(cur_result, f, indent=2) 20 | -------------------------------------------------------------------------------- /scripts/convert_seed_for_submission.py: -------------------------------------------------------------------------------- 1 | import os 2 | import json 3 | import argparse 4 | 5 | 6 | def get_args(): 7 | parser = argparse.ArgumentParser() 8 | parser.add_argument("--annotation-file", type=str) 9 | parser.add_argument("--result-file", type=str) 10 | parser.add_argument("--result-upload-file", type=str) 11 | return parser.parse_args() 12 | 13 | 14 | def eval_single(result_file, eval_only_type=None): 15 | results = {} 16 | for line in open(result_file): 17 | row = json.loads(line) 18 | results[row['question_id']] = row 19 | 20 | type_counts = {} 21 | correct_counts = {} 22 | for question_data in data['questions']: 23 | if eval_only_type is not None and question_data['data_type'] != eval_only_type: continue 24 | data_type = question_data['question_type_id'] 25 | type_counts[data_type] = type_counts.get(data_type, 0) + 1 26 | try: 27 | question_id = int(question_data['question_id']) 28 | except: 29 | question_id = question_data['question_id'] 30 | if question_id not in results: 31 | correct_counts[data_type] = correct_counts.get(data_type, 0) 32 | continue 33 | row = results[question_id] 34 | if row['text'] == question_data['answer']: 35 | correct_counts[data_type] = correct_counts.get(data_type, 0) + 1 36 | 37 | total_count = 0 38 | total_correct = 0 39 | for data_type in sorted(type_counts.keys()): 40 | accuracy = correct_counts[data_type] / type_counts[data_type] * 100 41 | if eval_only_type is None: 42 | print(f"{ques_type_id_to_name[data_type]}: {accuracy:.2f}%") 43 | 44 | total_count += type_counts[data_type] 45 | total_correct += correct_counts[data_type] 46 | 47 | total_accuracy = total_correct / total_count * 100 48 | if eval_only_type is None: 49 | print(f"Total accuracy: {total_accuracy:.2f}%") 50 | else: 51 | print(f"{eval_only_type} accuracy: {total_accuracy:.2f}%") 52 | 53 | return results 54 | 55 | if __name__ == "__main__": 56 | args = get_args() 57 | data = json.load(open(args.annotation_file)) 58 | ques_type_id_to_name = {id:n for n,id in data['question_type'].items()} 59 | 60 | results = eval_single(args.result_file) 61 | eval_single(args.result_file, eval_only_type='image') 62 | eval_single(args.result_file, eval_only_type='video') 63 | 64 | with open(args.result_upload_file, 'w') as fp: 65 | for question in data['questions']: 66 | qid = question['question_id'] 67 | if qid in results: 68 | result = results[qid] 69 | else: 70 | result = results[int(qid)] 71 | fp.write(json.dumps({ 72 | 'question_id': qid, 73 | 'prediction': result['text'] 74 | }) + '\n') 75 | -------------------------------------------------------------------------------- /scripts/convert_sqa_to_llava.py: -------------------------------------------------------------------------------- 1 | import json 2 | import os 3 | import fire 4 | import re 5 | from convert_sqa_to_llava_base_prompt import build_prompt_chatbot 6 | 7 | 8 | def convert_to_llava(base_dir, split, prompt_format="QCM-LEA"): 9 | split_indices = json.load(open(os.path.join(base_dir, "pid_splits.json")))[split] 10 | problems = json.load(open(os.path.join(base_dir, "problems.json"))) 11 | 12 | split_problems = build_prompt_chatbot( 13 | problems, split_indices, prompt_format, 14 | use_caption=False, is_test=False) 15 | 16 | target_format = [] 17 | for prob_id, (input, output) in split_problems.items(): 18 | if input.startswith('Question: '): 19 | input = input.replace('Question: ', '') 20 | if output.startswith('Answer: '): 21 | output = output.replace('Answer: ', '') 22 | 23 | raw_prob_data = problems[prob_id] 24 | if raw_prob_data['image'] is None: 25 | target_format.append({ 26 | "id": prob_id, 27 | "conversations": [ 28 | {'from': 'human', 'value': f"{input}"}, 29 | {'from': 'gpt', 'value': f"{output}"}, 30 | ], 31 | }) 32 | 33 | else: 34 | target_format.append({ 35 | "id": prob_id, 36 | "image": os.path.join(prob_id, raw_prob_data['image']), 37 | "conversations": [ 38 | {'from': 'human', 'value': f"{input}\n"}, 39 | {'from': 'gpt', 'value': f"{output}"}, 40 | ], 41 | }) 42 | 43 | print(f'Number of samples: {len(target_format)}') 44 | 45 | with open(os.path.join(base_dir, f"llava_{split}_{prompt_format}.json"), "w") as f: 46 | json.dump(target_format, f, indent=2) 47 | 48 | 49 | def convert_to_jsonl(base_dir, split, prompt_format="QCM-LEPA"): 50 | split_indices = json.load(open(os.path.join(base_dir, "pid_splits.json")))[split] 51 | problems = json.load(open(os.path.join(base_dir, "problems.json"))) 52 | 53 | split_problems = build_prompt_chatbot( 54 | problems, split_indices, prompt_format, 55 | use_caption=False, is_test=False) 56 | 57 | writer = open(os.path.join(base_dir, f"scienceqa_{split}_{prompt_format}.jsonl"), "w") 58 | for prob_id, (input, output) in split_problems.items(): 59 | if input.startswith('Question: '): 60 | input = input.replace('Question: ', '') 61 | if output.startswith('Answer: '): 62 | output = output.replace('Answer: ', '') 63 | 64 | raw_prob_data = problems[prob_id] 65 | if raw_prob_data['image'] is None: 66 | data = { 67 | "id": prob_id, 68 | "instruction": f"{input}", 69 | "output": f"{output}", 70 | } 71 | 72 | else: 73 | data = { 74 | "id": prob_id, 75 | "image": os.path.join(prob_id, raw_prob_data['image']), 76 | "instruction": f"{input}\n", 77 | "output": f"{output}", 78 | } 79 | writer.write(json.dumps(data) + '\n') 80 | writer.close() 81 | 82 | 83 | def main(task, **kwargs): 84 | globals()[task](**kwargs) 85 | 86 | 87 | if __name__ == "__main__": 88 | fire.Fire(main) 89 | -------------------------------------------------------------------------------- /scripts/convert_vizwiz_for_submission.py: -------------------------------------------------------------------------------- 1 | import os 2 | import argparse 3 | import json 4 | 5 | from llava.eval.m4c_evaluator import EvalAIAnswerProcessor 6 | 7 | 8 | def parse_args(): 9 | parser = argparse.ArgumentParser() 10 | parser.add_argument('--annotation-file', type=str, required=True) 11 | parser.add_argument('--result-file', type=str, required=True) 12 | parser.add_argument('--result-upload-file', type=str, required=True) 13 | return parser.parse_args() 14 | 15 | 16 | if __name__ == '__main__': 17 | 18 | args = parse_args() 19 | 20 | os.makedirs(os.path.dirname(args.result_upload_file), exist_ok=True) 21 | 22 | results = [] 23 | error_line = 0 24 | for line_idx, line in enumerate(open(args.result_file)): 25 | try: 26 | results.append(json.loads(line)) 27 | except: 28 | error_line += 1 29 | results = {x['question_id']: x['text'] for x in results} 30 | test_split = [json.loads(line) for line in open(args.annotation_file)] 31 | split_ids = set([x['question_id'] for x in test_split]) 32 | 33 | print(f'total results: {len(results)}, total split: {len(test_split)}, error_line: {error_line}') 34 | 35 | all_answers = [] 36 | 37 | answer_processor = EvalAIAnswerProcessor() 38 | 39 | for x in test_split: 40 | assert x['question_id'] in results 41 | all_answers.append({ 42 | 'image': x['image'], 43 | 'answer': answer_processor(results[x['question_id']]) 44 | }) 45 | 46 | with open(args.result_upload_file, 'w') as f: 47 | json.dump(all_answers, f) 48 | -------------------------------------------------------------------------------- /scripts/convert_vqav2_for_submission.py: -------------------------------------------------------------------------------- 1 | import os 2 | import argparse 3 | import json 4 | 5 | from llava.eval.m4c_evaluator import EvalAIAnswerProcessor 6 | 7 | 8 | def parse_args(): 9 | parser = argparse.ArgumentParser() 10 | parser.add_argument('--dir', type=str, default="./playground/data/eval/vqav2") 11 | parser.add_argument('--ckpt', type=str, required=True) 12 | parser.add_argument('--split', type=str, required=True) 13 | return parser.parse_args() 14 | 15 | 16 | if __name__ == '__main__': 17 | 18 | args = parse_args() 19 | 20 | src = os.path.join(args.dir, 'answers', args.split, args.ckpt, 'merge.jsonl') 21 | test_split = os.path.join(args.dir, 'llava_vqav2_mscoco_test2015.jsonl') 22 | dst = os.path.join(args.dir, 'answers_upload', args.split, f'{args.ckpt}.json') 23 | os.makedirs(os.path.dirname(dst), exist_ok=True) 24 | 25 | results = [] 26 | error_line = 0 27 | for line_idx, line in enumerate(open(src)): 28 | try: 29 | results.append(json.loads(line)) 30 | except: 31 | error_line += 1 32 | 33 | results = {x['question_id']: x['text'] for x in results} 34 | test_split = [json.loads(line) for line in open(test_split)] 35 | split_ids = set([x['question_id'] for x in test_split]) 36 | 37 | print(f'total results: {len(results)}, total split: {len(test_split)}, error_line: {error_line}') 38 | 39 | all_answers = [] 40 | 41 | answer_processor = EvalAIAnswerProcessor() 42 | 43 | for x in test_split: 44 | if x['question_id'] not in results: 45 | all_answers.append({ 46 | 'question_id': x['question_id'], 47 | 'answer': '' 48 | }) 49 | else: 50 | all_answers.append({ 51 | 'question_id': x['question_id'], 52 | 'answer': answer_processor(results[x['question_id']]) 53 | }) 54 | 55 | with open(dst, 'w') as f: 56 | json.dump(all_answers, open(dst, 'w')) 57 | -------------------------------------------------------------------------------- /scripts/eval_mmvet.h: -------------------------------------------------------------------------------- 1 | bash scripts/v1_5/eval/eval_multi_lora.sh ../pretrained_weights/llava1.5_7b ./checkpoints/dpo/llava1.5_7b-lora32-lr2e-6-shargpt4_3w_llava_3w_coco_3w-1e playground/data/eval/mm-vet.jsonl results/mmvet/llava1.5_7b-lora32-lr2e-6-shargpt4_3w_llava_3w_coco_3w-1e playground/data/eval/mmvet_images 8 0 0 2 | bash scripts/v1_5/eval/eval_multi_lora.sh ../pretrained_weights/llava1.5_7b ./checkpoints/dpo/llava1.5_7b-lora32-lr2e-6-shargpt4_4w_llava_4w_coco_4w-1e playground/data/eval/mm-vet.jsonl results/mmvet/llava1.5_7b-lora32-lr2e-6-shargpt4_4w_llava_4w_coco_4w-1e playground/data/eval/mmvet_images 8 0 0 3 | bash scripts/v1_5/eval/eval_multi_lora.sh ../pretrained_weights/llava1.5_7b ./checkpoints/dpo/llava1.5_7b-lora32-lr2e-6-shargpt4_5w_llava_5w_coco_5w-1e playground/data/eval/mm-vet.jsonl results/mmvet/llava1.5_7b-lora32-lr2e-6-shargpt4_5w_llava_5w_coco_5w-1e playground/data/eval/mmvet_images 8 0 0 4 | bash scripts/v1_5/eval/eval_multi_lora.sh ../pretrained_weights/llava1.5_7b ./checkpoints/dpo/llava1.5_7b-lora32-lr2e-6-shargpt4_6w_llava_6w_coco_6w-1e playground/data/eval/mm-vet.jsonl results/mmvet/llava1.5_7b-lora32-lr2e-6-shargpt4_6w_llava_6w_coco_6w-1e playground/data/eval/mmvet_images 8 0 0 5 | -------------------------------------------------------------------------------- /scripts/extract_mm_projector.py: -------------------------------------------------------------------------------- 1 | """ 2 | This is just a utility that I use to extract the projector for quantized models. 3 | It is NOT necessary at all to train, or run inference/serve demos. 4 | Use this script ONLY if you fully understand its implications. 5 | """ 6 | 7 | 8 | import os 9 | import argparse 10 | import torch 11 | import json 12 | from collections import defaultdict 13 | 14 | 15 | def parse_args(): 16 | parser = argparse.ArgumentParser(description='Extract MMProjector weights') 17 | parser.add_argument('--model-path', type=str, help='model folder') 18 | parser.add_argument('--output', type=str, help='output file') 19 | args = parser.parse_args() 20 | return args 21 | 22 | 23 | if __name__ == '__main__': 24 | args = parse_args() 25 | 26 | keys_to_match = ['mm_projector'] 27 | ckpt_to_key = defaultdict(list) 28 | try: 29 | model_indices = json.load(open(os.path.join(args.model_path, 'pytorch_model.bin.index.json'))) 30 | for k, v in model_indices['weight_map'].items(): 31 | if any(key_match in k for key_match in keys_to_match): 32 | ckpt_to_key[v].append(k) 33 | except FileNotFoundError: 34 | # Smaller models or model checkpoints saved by DeepSpeed. 35 | v = 'pytorch_model.bin' 36 | for k in torch.load(os.path.join(args.model_path, v), map_location='cpu').keys(): 37 | if any(key_match in k for key_match in keys_to_match): 38 | ckpt_to_key[v].append(k) 39 | 40 | loaded_weights = {} 41 | 42 | for ckpt_name, weight_keys in ckpt_to_key.items(): 43 | ckpt = torch.load(os.path.join(args.model_path, ckpt_name), map_location='cpu') 44 | for k in weight_keys: 45 | loaded_weights[k] = ckpt[k] 46 | 47 | torch.save(loaded_weights, args.output) 48 | -------------------------------------------------------------------------------- /scripts/finetune.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # IMPORTANT: this is the training script for the original LLaVA, NOT FOR LLaVA V1.5! 4 | 5 | # Uncomment and set the following variables correspondingly to run this script: 6 | 7 | ################## VICUNA ################## 8 | # PROMPT_VERSION=v1 9 | # MODEL_VERSION="vicuna-v1-3-7b" 10 | ################## VICUNA ################## 11 | 12 | ################## LLaMA-2 ################## 13 | # PROMPT_VERSION="llava_llama_2" 14 | # MODEL_VERSION="llama-2-7b-chat" 15 | ################## LLaMA-2 ################## 16 | 17 | deepspeed llava/train/train_mem.py \ 18 | --deepspeed ./scripts/zero2.json \ 19 | --model_name_or_path ./checkpoints/$MODEL_VERSION \ 20 | --version $PROMPT_VERSION \ 21 | --data_path ./playground/data/llava_instruct_80k.json \ 22 | --image_folder /path/to/coco/train2017 \ 23 | --vision_tower openai/clip-vit-large-patch14 \ 24 | --pretrain_mm_mlp_adapter ./checkpoints/llava-$MODEL_VERSION-pretrain/mm_projector.bin \ 25 | --mm_vision_select_layer -2 \ 26 | --mm_use_im_start_end False \ 27 | --mm_use_im_patch_token False \ 28 | --bf16 True \ 29 | --output_dir ./checkpoints/llava-$MODEL_VERSION-finetune \ 30 | --num_train_epochs 1 \ 31 | --per_device_train_batch_size 16 \ 32 | --per_device_eval_batch_size 4 \ 33 | --gradient_accumulation_steps 1 \ 34 | --evaluation_strategy "no" \ 35 | --save_strategy "steps" \ 36 | --save_steps 50000 \ 37 | --save_total_limit 1 \ 38 | --learning_rate 2e-5 \ 39 | --weight_decay 0. \ 40 | --warmup_ratio 0.03 \ 41 | --lr_scheduler_type "cosine" \ 42 | --logging_steps 1 \ 43 | --tf32 True \ 44 | --model_max_length 2048 \ 45 | --gradient_checkpointing True \ 46 | --dataloader_num_workers 4 \ 47 | --lazy_preprocess True \ 48 | --report_to wandb 49 | -------------------------------------------------------------------------------- /scripts/finetune_bpo.sh: -------------------------------------------------------------------------------- 1 | deepspeed llava/train/bpo_llava.py \ 2 | --mm_projector_lr 2e-6 \ 3 | --mm_projector_type mlp2x_gelu \ 4 | --learning_rate 2e-6 \ 5 | --deepspeed ./scripts/zero2.json \ 6 | --lora_enable True \ 7 | --lora_r 32 \ 8 | --lora_alpha 256 \ 9 | --model_name_or_path path-to-model \ 10 | --version v1 \ 11 | --data_path path-to-json-annotation-file \ 12 | --image_folder path-to-image-folder \ 13 | --vision_tower openai/clip-vit-large-patch14 \ 14 | --mm_vision_select_layer -2 \ 15 | --mm_use_im_start_end False \ 16 | --mm_use_im_patch_token False \ 17 | --bf16 True \ 18 | --output_dir path-to-output \ 19 | --num_train_epochs 2 \ 20 | --per_device_train_batch_size 4 \ 21 | --per_device_eval_batch_size 4 \ 22 | --gradient_accumulation_steps 1 \ 23 | --evaluation_strategy "no" \ 24 | --save_strategy "no" \ 25 | --save_steps 5000 \ 26 | --save_total_limit 1 \ 27 | --weight_decay 0. \ 28 | --warmup_ratio 0.03 \ 29 | --lr_scheduler_type "cosine" \ 30 | --logging_steps 1 \ 31 | --tf32 True \ 32 | --model_max_length 2048 \ 33 | --gradient_checkpointing True \ 34 | --dataloader_num_workers 4 \ 35 | --lazy_preprocess True \ 36 | --lora_enable 37 | 38 | bash scripts/v1_5/eval/eval_multi_lora.sh path-to-model path-to-lora playground/data/eval/mm-vet.jsonl path-to-result path-to-images gpu-num temperature start_gpu 39 | python scripts/convert_mmvet_for_eval.py --src path-to-result-jsonl --dst path-to-result-json 40 | -------------------------------------------------------------------------------- /scripts/finetune_bpo_flash.sh: -------------------------------------------------------------------------------- 1 | deepspeed llava/train/bpo_llava_flash.py \ 2 | --mm_projector_lr 2e-6 \ 3 | --mm_projector_type mlp2x_gelu \ 4 | --learning_rate 2e-6 \ 5 | --deepspeed ./scripts/zero2.json \ 6 | --lora_enable True \ 7 | --lora_r 32 \ 8 | --lora_alpha 256 \ 9 | --model_name_or_path path-to-model \ 10 | --version v1 \ 11 | --data_path path-to-json-annotation-file \ 12 | --image_folder path-to-image-folder \ 13 | --vision_tower openai/clip-vit-large-patch14 \ 14 | --mm_vision_select_layer -2 \ 15 | --mm_use_im_start_end False \ 16 | --mm_use_im_patch_token False \ 17 | --bf16 True \ 18 | --output_dir path-to-output \ 19 | --num_train_epochs 2 \ 20 | --per_device_train_batch_size 4 \ 21 | --per_device_eval_batch_size 4 \ 22 | --gradient_accumulation_steps 1 \ 23 | --evaluation_strategy "no" \ 24 | --save_strategy "no" \ 25 | --save_steps 5000 \ 26 | --save_total_limit 1 \ 27 | --weight_decay 0. \ 28 | --warmup_ratio 0.03 \ 29 | --lr_scheduler_type "cosine" \ 30 | --logging_steps 1 \ 31 | --tf32 True \ 32 | --model_max_length 2048 \ 33 | --gradient_checkpointing True \ 34 | --dataloader_num_workers 4 \ 35 | --lazy_preprocess True \ 36 | --lora_enable 37 | 38 | bash scripts/v1_5/eval/eval_multi_lora.sh path-to-model path-to-lora playground/data/eval/mm-vet.jsonl path-to-result path-to-images gpu-num temperature start_gpu 39 | python scripts/convert_mmvet_for_eval.py --src path-to-result-jsonl --dst path-to-result-json 40 | -------------------------------------------------------------------------------- /scripts/finetune_full_schedule.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # IMPORTANT: this is the training script for the original LLaVA, NOT FOR LLaVA V1.5! 4 | 5 | # Uncomment and set the following variables correspondingly to run this script: 6 | 7 | ################## VICUNA ################## 8 | # PROMPT_VERSION=v1 9 | # MODEL_VERSION="vicuna-v1-3-7b" 10 | ################## VICUNA ################## 11 | 12 | ################## LLaMA-2 ################## 13 | # PROMPT_VERSION="llava_llama_2" 14 | # MODEL_VERSION="llama-2-7b-chat" 15 | ################## LLaMA-2 ################## 16 | 17 | deepspeed llava/train/train_mem.py \ 18 | --deepspeed ./scripts/zero2.json \ 19 | --model_name_or_path ./checkpoints/$MODEL_VERSION \ 20 | --version $PROMPT_VERSION \ 21 | --data_path ./playground/data/llava_instruct_158k.json \ 22 | --image_folder /path/to/coco/train2017 \ 23 | --vision_tower openai/clip-vit-large-patch14 \ 24 | --pretrain_mm_mlp_adapter ./checkpoints/llava-$MODEL_VERSION-pretrain/mm_projector.bin \ 25 | --mm_vision_select_layer -2 \ 26 | --mm_use_im_start_end False \ 27 | --mm_use_im_patch_token False \ 28 | --bf16 True \ 29 | --output_dir ./checkpoints/llava-$MODEL_VERSION-finetune \ 30 | --num_train_epochs 3 \ 31 | --per_device_train_batch_size 16 \ 32 | --per_device_eval_batch_size 4 \ 33 | --gradient_accumulation_steps 1 \ 34 | --evaluation_strategy "no" \ 35 | --save_strategy "steps" \ 36 | --save_steps 50000 \ 37 | --save_total_limit 1 \ 38 | --learning_rate 2e-5 \ 39 | --weight_decay 0. \ 40 | --warmup_ratio 0.03 \ 41 | --lr_scheduler_type "cosine" \ 42 | --logging_steps 1 \ 43 | --tf32 True \ 44 | --model_max_length 2048 \ 45 | --gradient_checkpointing True \ 46 | --dataloader_num_workers 4 \ 47 | --lazy_preprocess True \ 48 | --report_to wandb 49 | -------------------------------------------------------------------------------- /scripts/finetune_lora.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # IMPORTANT: this is the training script for the original LLaVA, NOT FOR LLaVA V1.5! 4 | 5 | # Uncomment and set the following variables correspondingly to run this script: 6 | 7 | ################## VICUNA ################## 8 | # PROMPT_VERSION=v1 9 | # MODEL_VERSION="vicuna-v1-3-7b" 10 | ################## VICUNA ################## 11 | 12 | ################## LLaMA-2 ################## 13 | # PROMPT_VERSION="llava_llama_2" 14 | # MODEL_VERSION="llama-2-7b-chat" 15 | ################## LLaMA-2 ################## 16 | 17 | deepspeed llava/train/train_mem.py \ 18 | --deepspeed ./scripts/zero2.json \ 19 | --lora_enable True \ 20 | --model_name_or_path ./checkpoints/$MODEL_VERSION \ 21 | --version $PROMPT_VERSION \ 22 | --data_path ./playground/data/llava_instruct_80k.json \ 23 | --image_folder /path/to/coco/train2017 \ 24 | --vision_tower openai/clip-vit-large-patch14 \ 25 | --pretrain_mm_mlp_adapter ./checkpoints/llava-$MODEL_VERSION-pretrain/mm_projector.bin \ 26 | --mm_vision_select_layer -2 \ 27 | --mm_use_im_start_end False \ 28 | --mm_use_im_patch_token False \ 29 | --bf16 True \ 30 | --output_dir ./checkpoints/llava-$MODEL_VERSION-finetune_lora \ 31 | --num_train_epochs 1 \ 32 | --per_device_train_batch_size 16 \ 33 | --per_device_eval_batch_size 4 \ 34 | --gradient_accumulation_steps 1 \ 35 | --evaluation_strategy "no" \ 36 | --save_strategy "steps" \ 37 | --save_steps 50000 \ 38 | --save_total_limit 1 \ 39 | --learning_rate 2e-5 \ 40 | --weight_decay 0. \ 41 | --warmup_ratio 0.03 \ 42 | --lr_scheduler_type "cosine" \ 43 | --logging_steps 1 \ 44 | --tf32 True \ 45 | --model_max_length 2048 \ 46 | --gradient_checkpointing True \ 47 | --lazy_preprocess True \ 48 | --dataloader_num_workers 4 \ 49 | --report_to wandb 50 | -------------------------------------------------------------------------------- /scripts/finetune_qlora.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # IMPORTANT: this is the training script for the original LLaVA, NOT FOR LLaVA V1.5! 4 | 5 | # Uncomment and set the following variables correspondingly to run this script: 6 | 7 | ################## VICUNA ################## 8 | # PROMPT_VERSION=v1 9 | # MODEL_VERSION="vicuna-v1-3-7b" 10 | ################## VICUNA ################## 11 | 12 | ################## LLaMA-2 ################## 13 | # PROMPT_VERSION="llava_llama_2" 14 | # MODEL_VERSION="llama-2-7b-chat" 15 | ################## LLaMA-2 ################## 16 | 17 | deepspeed llava/train/train_mem.py \ 18 | --deepspeed ./scripts/zero2.json \ 19 | --lora_enable True \ 20 | --bits 4 \ 21 | --model_name_or_path ./checkpoints/$MODEL_VERSION \ 22 | --version $PROMPT_VERSION \ 23 | --data_path ./playground/data/llava_instruct_80k.json \ 24 | --image_folder /path/to/coco/train2017 \ 25 | --vision_tower openai/clip-vit-large-patch14 \ 26 | --pretrain_mm_mlp_adapter ./checkpoints/llava-$MODEL_VERSION-pretrain/mm_projector.bin \ 27 | --mm_vision_select_layer -2 \ 28 | --mm_use_im_start_end False \ 29 | --mm_use_im_patch_token False \ 30 | --bf16 True \ 31 | --output_dir ./checkpoints/llava-$MODEL_VERSION-finetune_lora \ 32 | --num_train_epochs 1 \ 33 | --per_device_train_batch_size 16 \ 34 | --per_device_eval_batch_size 4 \ 35 | --gradient_accumulation_steps 1 \ 36 | --evaluation_strategy "no" \ 37 | --save_strategy "steps" \ 38 | --save_steps 50000 \ 39 | --save_total_limit 1 \ 40 | --learning_rate 2e-5 \ 41 | --weight_decay 0. \ 42 | --warmup_ratio 0.03 \ 43 | --lr_scheduler_type "cosine" \ 44 | --logging_steps 1 \ 45 | --tf32 True \ 46 | --model_max_length 2048 \ 47 | --gradient_checkpointing True \ 48 | --lazy_preprocess True \ 49 | --dataloader_num_workers 4 \ 50 | --report_to wandb 51 | -------------------------------------------------------------------------------- /scripts/finetune_sft.sh: -------------------------------------------------------------------------------- 1 | deepspeed --include=localhost:0,1,2,3,4,5,6,7 llava/train/train.py --mm_projector_lr 2e-6 --mm_projector_type mlp2x_gelu --learning_rate 2e-6 --deepspeed ./scripts/zero2.json --lora_enable True --lora_r 32 --lora_alpha 256 --model_name_or_path ../pretrained_weights/llava1.5_7b --version v1 --data_path playground/data/train/dpo/sft/vicuna1.3_shargpt4_1w_llavar_whole_coco3w.json --image_folder ../data/sharegpt4v/images --vision_tower openai/clip-vit-large-patch14 --pretrain_mm_mlp_adapter ../pretrained_weights/llava1.5_7b/mm_projector.bin --mm_vision_select_layer -2 --mm_use_im_start_end False --mm_use_im_patch_token False --bf16 True --output_dir ./checkpoints/dpo/sft/llava1.5_7b-lora32-lr2e-6-shargpt4_1w_llavar_whole_coco3w-1e --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 1 --evaluation_strategy "no" --save_strategy "no" --save_steps 50000000 --save_total_limit 1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type "cosine" --logging_steps 1 --tf32 True --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 4 --lazy_preprocess True --lora_enable 2 | deepspeed --include=localhost:0,1,2,3,4,5,6,7 llava/train/train.py --mm_projector_lr 1e-5 --mm_projector_type mlp2x_gelu --learning_rate 1e-5 --deepspeed ./scripts/zero2.json --lora_enable True --lora_r 32 --lora_alpha 256 --model_name_or_path ../pretrained_weights/llava1.5_7b --version v1 --data_path playground/data/train/dpo/sft/vicuna1.3_shargpt4_1w_llavar_whole_coco3w.json --image_folder ../data/sharegpt4v/images --vision_tower openai/clip-vit-large-patch14 --pretrain_mm_mlp_adapter ../pretrained_weights/llava1.5_7b/mm_projector.bin --mm_vision_select_layer -2 --mm_use_im_start_end False --mm_use_im_patch_token False --bf16 True --output_dir ./checkpoints/dpo/sft/llava1.5_7b-lora32-lr1e-5-shargpt4_1w_llavar_whole_coco3w-1e --num_train_epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 1 --evaluation_strategy "no" --save_strategy "no" --save_steps 50000000 --save_total_limit 1 --weight_decay 0. --warmup_ratio 0.03 --lr_scheduler_type "cosine" --logging_steps 1 --tf32 True --model_max_length 2048 --gradient_checkpointing True --dataloader_num_workers 4 --lazy_preprocess True --lora_enable 3 | -------------------------------------------------------------------------------- /scripts/finetune_sqa.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # IMPORTANT: this is the training script for the original LLaVA, NOT FOR LLaVA V1.5! 4 | 5 | deepspeed llava/train/train_mem.py \ 6 | --deepspeed ./scripts/zero2.json \ 7 | --model_name_or_path lmsys/vicuna-13b-v1.3 \ 8 | --version $PROMPT_VERSION \ 9 | --data_path /Data/ScienceQA/data/scienceqa/llava_train_QCM-LEA.json \ 10 | --image_folder /Data/ScienceQA/data/scienceqa/images/train \ 11 | --vision_tower openai/clip-vit-large-patch14 \ 12 | --pretrain_mm_mlp_adapter ./checkpoints/huggingface/liuhaotian/llava-pretrain-vicuna-13b-v1.3/mm_projector.bin \ 13 | --mm_vision_select_layer -2 \ 14 | --mm_use_im_start_end False \ 15 | --mm_use_im_patch_token False \ 16 | --bf16 True \ 17 | --output_dir ./checkpoints/llava-vicuna-13b-v1.3-pretrain_lcs558k_plain-ScienceQA_QCM_LEA-12e \ 18 | --num_train_epochs 12 \ 19 | --per_device_train_batch_size 16 \ 20 | --per_device_eval_batch_size 4 \ 21 | --gradient_accumulation_steps 1 \ 22 | --evaluation_strategy "no" \ 23 | --save_strategy "steps" \ 24 | --save_steps 50000 \ 25 | --save_total_limit 1 \ 26 | --learning_rate 2e-5 \ 27 | --weight_decay 0. \ 28 | --warmup_ratio 0.03 \ 29 | --lr_scheduler_type "cosine" \ 30 | --logging_steps 1 \ 31 | --tf32 True \ 32 | --model_max_length 2048 \ 33 | --gradient_checkpointing True \ 34 | --dataloader_num_workers 4 \ 35 | --lazy_preprocess True \ 36 | --report_to wandb 37 | -------------------------------------------------------------------------------- /scripts/merge_lora_weights.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | from llava.model.builder import load_pretrained_model 3 | from llava.mm_utils import get_model_name_from_path 4 | 5 | 6 | def merge_lora(args): 7 | model_name = get_model_name_from_path(args.model_path) 8 | tokenizer, model, image_processor, context_len = load_pretrained_model(args.model_path, args.model_base, model_name, device_map='cpu') 9 | 10 | model.save_pretrained(args.save_model_path) 11 | tokenizer.save_pretrained(args.save_model_path) 12 | 13 | 14 | if __name__ == "__main__": 15 | parser = argparse.ArgumentParser() 16 | parser.add_argument("--model-path", type=str, required=True) 17 | parser.add_argument("--model-base", type=str, required=True) 18 | parser.add_argument("--save-model-path", type=str, required=True) 19 | 20 | args = parser.parse_args() 21 | 22 | merge_lora(args) 23 | -------------------------------------------------------------------------------- /scripts/merge_lora_weights_reward.py: -------------------------------------------------------------------------------- 1 | import argparse 2 | from llava.model.builder import load_pretrained_model, load_reward_model 3 | from llava.mm_utils import get_model_name_from_path 4 | 5 | 6 | def merge_lora(args): 7 | model_name = get_model_name_from_path(args.model_path) 8 | tokenizer, model, image_processor, context_len = load_reward_model(args.model_path, args.model_base, model_name, device_map='cpu') 9 | 10 | model.save_pretrained(args.save_model_path) 11 | tokenizer.save_pretrained(args.save_model_path) 12 | 13 | 14 | if __name__ == "__main__": 15 | parser = argparse.ArgumentParser() 16 | parser.add_argument("--model-path", type=str, required=True) 17 | parser.add_argument("--model-base", type=str, required=True) 18 | parser.add_argument("--save-model-path", type=str, required=True) 19 | 20 | args = parser.parse_args() 21 | 22 | merge_lora(args) 23 | -------------------------------------------------------------------------------- /scripts/pretrain.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # IMPORTANT: this is the training script for the original LLaVA, NOT FOR LLaVA V1.5! 4 | 5 | # Uncomment and set the following variables correspondingly to run this script: 6 | 7 | # MODEL_VERSION=vicuna-v1-3-7b 8 | # MODEL_VERSION=llama-2-7b-chat 9 | 10 | ########### DO NOT CHANGE ########### 11 | ########### USE THIS FOR BOTH ########### 12 | PROMPT_VERSION=plain 13 | ########### DO NOT CHANGE ########### 14 | 15 | deepspeed llava/train/train_mem.py \ 16 | --deepspeed ./scripts/zero2.json \ 17 | --model_name_or_path ./checkpoints/$MODEL_VERSION \ 18 | --version $PROMPT_VERSION \ 19 | --data_path /path/to/pretrain_data.json \ 20 | --image_folder /path/to/images \ 21 | --vision_tower openai/clip-vit-large-patch14 \ 22 | --tune_mm_mlp_adapter True \ 23 | --mm_vision_select_layer -2 \ 24 | --mm_use_im_start_end False \ 25 | --mm_use_im_patch_token False \ 26 | --bf16 True \ 27 | --output_dir ./checkpoints/llava-$MODEL_VERSION-pretrain \ 28 | --num_train_epochs 1 \ 29 | --per_device_train_batch_size 16 \ 30 | --per_device_eval_batch_size 4 \ 31 | --gradient_accumulation_steps 1 \ 32 | --evaluation_strategy "no" \ 33 | --save_strategy "steps" \ 34 | --save_steps 24000 \ 35 | --save_total_limit 1 \ 36 | --learning_rate 2e-3 \ 37 | --weight_decay 0. \ 38 | --warmup_ratio 0.03 \ 39 | --lr_scheduler_type "cosine" \ 40 | --logging_steps 1 \ 41 | --tf32 True \ 42 | --model_max_length 2048 \ 43 | --gradient_checkpointing True \ 44 | --dataloader_num_workers 4 \ 45 | --lazy_preprocess True \ 46 | --report_to wandb 47 | -------------------------------------------------------------------------------- /scripts/pretrain_xformers.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # Uncomment and set the following variables correspondingly to run this script: 4 | 5 | # MODEL_VERSION=vicuna-v1-3-7b 6 | # MODEL_VERSION=llama-2-7b-chat 7 | 8 | ########### DO NOT CHANGE ########### 9 | ########### USE THIS FOR BOTH ########### 10 | PROMPT_VERSION=plain 11 | ########### DO NOT CHANGE ########### 12 | 13 | deepspeed llava/train/train_xformers.py \ 14 | --deepspeed ./scripts/zero2.json \ 15 | --model_name_or_path ./checkpoints/$MODEL_VERSION \ 16 | --version $PROMPT_VERSION \ 17 | --data_path /path/to/pretrain_data.json \ 18 | --image_folder /path/to/images \ 19 | --vision_tower openai/clip-vit-large-patch14 \ 20 | --tune_mm_mlp_adapter True \ 21 | --mm_vision_select_layer -2 \ 22 | --mm_use_im_start_end False \ 23 | --mm_use_im_patch_token False \ 24 | --bf16 False \ 25 | --output_dir ./checkpoints/llava-$MODEL_VERSION-pretrain \ 26 | --num_train_epochs 1 \ 27 | --per_device_train_batch_size 4 \ 28 | --per_device_eval_batch_size 4 \ 29 | --gradient_accumulation_steps 4 \ 30 | --evaluation_strategy "no" \ 31 | --save_strategy "steps" \ 32 | --save_steps 24000 \ 33 | --save_total_limit 1 \ 34 | --learning_rate 2e-3 \ 35 | --weight_decay 0. \ 36 | --warmup_ratio 0.03 \ 37 | --lr_scheduler_type "cosine" \ 38 | --logging_steps 1 \ 39 | --tf32 False \ 40 | --model_max_length 2048 \ 41 | --gradient_checkpointing True \ 42 | --dataloader_num_workers 4 \ 43 | --lazy_preprocess True \ 44 | --report_to wandb 45 | -------------------------------------------------------------------------------- /scripts/sqa_eval_batch.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | CHUNKS=8 4 | for IDX in {0..7}; do 5 | CUDA_VISIBLE_DEVICES=$IDX python -m llava.eval.model_vqa_science \ 6 | --model-path liuhaotian/llava-lcs558k-scienceqa-vicuna-13b-v1.3 \ 7 | --question-file ~/haotian/datasets/ScienceQA/data/scienceqa/llava_test_QCM-LEA.json \ 8 | --image-folder ~/haotian/datasets/ScienceQA/data/scienceqa/images/test \ 9 | --answers-file ./test_llava-13b-chunk$CHUNKS_$IDX.jsonl \ 10 | --num-chunks $CHUNKS \ 11 | --chunk-idx $IDX \ 12 | --conv-mode llava_v1 & 13 | done 14 | -------------------------------------------------------------------------------- /scripts/sqa_eval_gather.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | CHUNKS=8 4 | output_file="test_llava-13b.jsonl" 5 | 6 | # Clear out the output file if it exists. 7 | > "$output_file" 8 | 9 | # Loop through the indices and concatenate each file. 10 | for idx in $(seq 0 $((CHUNKS-1))); do 11 | cat "./test_llava-13b-chunk${idx}.jsonl" >> "$output_file" 12 | done 13 | 14 | python llava/eval/eval_science_qa.py \ 15 | --base-dir ~/haotian/datasets/ScienceQA/data/scienceqa \ 16 | --result-file ./test_llava-13b.jsonl \ 17 | --output-file ./test_llava-13b_output.json \ 18 | --output-result ./test_llava-13b_result.json 19 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/bootstrap_dpo_multi.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # Check if three arguments are passed 4 | if [ "$#" -ne 10 ]; then 5 | echo "Usage: $0 " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_base=$1 11 | model_path=$2 12 | question_path=$3 13 | base_answer_path=$4 14 | image_folder=$5 15 | num_samples=$6 16 | subset=$7 17 | N=$8 18 | temperature=$9 19 | GS=${10} 20 | 21 | # Loop over each chunk/process 22 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 47 | for ((i=0; i> "${base_answer_path}_merged.jsonl" 50 | done 51 | # remove the unmerged files 52 | for (( chunk_id=0; chunk_id " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_base=$1 11 | model_path=$2 12 | question_path=$3 13 | base_answer_path=$4 14 | image_folder=$5 15 | num_samples=$6 16 | subset=$7 17 | N=$8 18 | temperature=$9 19 | GS=${10} 20 | 21 | # Loop over each chunk/process 22 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 47 | for ((i=0; i> "${base_answer_path}_merged.jsonl" 50 | done 51 | # remove the unmerged files 52 | for (( chunk_id=0; chunk_id " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_path=$1 11 | question_path=$2 12 | base_answer_path=$3 13 | image_folder=$4 14 | N=$5 15 | temperature=$6 16 | GS=$7 17 | 18 | # Loop over each chunk/process 19 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 43 | for ((i=0; i> "${base_answer_path}_merged.jsonl" 46 | done 47 | # remove the unmerged files 48 | for (( chunk_id=0; chunk_id " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_path=$1 11 | lora_path=$2 12 | question_path=$3 13 | base_answer_path=$4 14 | image_folder=$5 15 | N=$6 16 | temperature=$7 17 | GS=$8 18 | 19 | # Loop over each chunk/process 20 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 44 | for ((i=0; i> "${base_answer_path}_merged.jsonl" 47 | done 48 | # remove the unmerged files 49 | for (( chunk_id=0; chunk_id " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_path=$1 11 | question_path=$2 12 | base_answer_path=$3 13 | image_folder=$4 14 | N=$5 15 | temperature=$6 16 | GS=$7 17 | TYPE=$8 18 | 19 | # Loop over each chunk/process 20 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 44 | for ((i=0; i> "${base_answer_path}_merged.jsonl" 47 | done 48 | # remove the unmerged files 49 | for (( chunk_id=0; chunk_id " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_path=$1 11 | lora_path=$2 12 | question_path=$3 13 | base_answer_path=$4 14 | image_folder=$5 15 | N=$6 16 | temperature=$7 17 | GS=$8 18 | 19 | # Loop over each chunk/process 20 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 44 | for ((i=0; i> "${base_answer_path}.jsonl" 47 | done 48 | # remove the unmerged files 49 | for (( chunk_id=0; chunk_id " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_path=$1 11 | question_path=$2 12 | base_answer_path=$3 13 | image_folder=$4 14 | N=$5 15 | temperature=$6 16 | GS=$7 17 | 18 | # Loop over each chunk/process 19 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 43 | for ((i=0; i> "${base_answer_path}.jsonl" 46 | done 47 | # remove the unmerged files 48 | for (( chunk_id=0; chunk_id "$output_file" 30 | 31 | # Loop through the indices and concatenate each file. 32 | for IDX in $(seq 0 $((CHUNKS-1))); do 33 | cat ./playground/data/eval/gqa/answers/$SPLIT/$CKPT/${CHUNKS}_${IDX}.jsonl >> "$output_file" 34 | done 35 | 36 | python scripts/convert_gqa_for_eval.py --src $output_file --dst $GQADIR/testdev_balanced_predictions.json 37 | 38 | cd $GQADIR 39 | python eval/eval.py --tier testdev_balanced 40 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/llavabench.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | python -m llava.eval.model_vqa \ 4 | --model-path liuhaotian/llava-v1.5-13b \ 5 | --question-file ./playground/data/eval/llava-bench-in-the-wild/questions.jsonl \ 6 | --image-folder ./playground/data/eval/llava-bench-in-the-wild/images \ 7 | --answers-file ./playground/data/eval/llava-bench-in-the-wild/answers/llava-v1.5-13b.jsonl \ 8 | --temperature 0 \ 9 | --conv-mode vicuna_v1 10 | 11 | mkdir -p playground/data/eval/llava-bench-in-the-wild/reviews 12 | 13 | python llava/eval/eval_gpt_review_bench.py \ 14 | --question playground/data/eval/llava-bench-in-the-wild/questions.jsonl \ 15 | --context playground/data/eval/llava-bench-in-the-wild/context.jsonl \ 16 | --rule llava/eval/table/rule.json \ 17 | --answer-list \ 18 | playground/data/eval/llava-bench-in-the-wild/answers_gpt4.jsonl \ 19 | playground/data/eval/llava-bench-in-the-wild/answers/llava-v1.5-13b.jsonl \ 20 | --output \ 21 | playground/data/eval/llava-bench-in-the-wild/reviews/llava-v1.5-13b.jsonl 22 | 23 | python llava/eval/summarize_gpt_review.py -f playground/data/eval/llava-bench-in-the-wild/reviews/llava-v1.5-13b.jsonl 24 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/mmbench.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | SPLIT="mmbench_dev_20230712" 4 | 5 | python -m llava.eval.model_vqa_mmbench \ 6 | --model-path liuhaotian/llava-v1.5-13b \ 7 | --question-file ./playground/data/eval/mmbench/$SPLIT.tsv \ 8 | --answers-file ./playground/data/eval/mmbench/answers/$SPLIT/llava-v1.5-13b.jsonl \ 9 | --single-pred-prompt \ 10 | --temperature 0 \ 11 | --conv-mode vicuna_v1 12 | 13 | mkdir -p playground/data/eval/mmbench/answers_upload/$SPLIT 14 | 15 | python scripts/convert_mmbench_for_submission.py \ 16 | --annotation-file ./playground/data/eval/mmbench/$SPLIT.tsv \ 17 | --result-dir ./playground/data/eval/mmbench/answers/$SPLIT \ 18 | --upload-dir ./playground/data/eval/mmbench/answers_upload/$SPLIT \ 19 | --experiment llava-v1.5-13b 20 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/mmbench_cn.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | SPLIT="mmbench_dev_cn_20231003" 4 | 5 | python -m llava.eval.model_vqa_mmbench \ 6 | --model-path liuhaotian/llava-v1.5-13b \ 7 | --question-file ./playground/data/eval/mmbench_cn/$SPLIT.tsv \ 8 | --answers-file ./playground/data/eval/mmbench_cn/answers/$SPLIT/llava-v1.5-13b.jsonl \ 9 | --lang cn \ 10 | --single-pred-prompt \ 11 | --temperature 0 \ 12 | --conv-mode vicuna_v1 13 | 14 | mkdir -p playground/data/eval/mmbench/answers_upload/$SPLIT 15 | 16 | python scripts/convert_mmbench_for_submission.py \ 17 | --annotation-file ./playground/data/eval/mmbench_cn/$SPLIT.tsv \ 18 | --result-dir ./playground/data/eval/mmbench_cn/answers/$SPLIT \ 19 | --upload-dir ./playground/data/eval/mmbench_cn/answers_upload/$SPLIT \ 20 | --experiment llava-v1.5-13b 21 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/mme.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | python -m llava.eval.model_vqa_loader \ 4 | --model-path liuhaotian/llava-v1.5-13b \ 5 | --question-file ./playground/data/eval/MME/llava_mme.jsonl \ 6 | --image-folder ./playground/data/eval/MME/MME_Benchmark_release_version \ 7 | --answers-file ./playground/data/eval/MME/answers/llava-v1.5-13b.jsonl \ 8 | --temperature 0 \ 9 | --conv-mode vicuna_v1 10 | 11 | cd ./playground/data/eval/MME 12 | 13 | python convert_answer_to_mme.py --experiment llava-v1.5-13b 14 | 15 | cd eval_tool 16 | 17 | python calculation.py --results_dir answers/llava-v1.5-13b 18 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/mmvet.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | python -m llava.eval.model_vqa \ 4 | --model-path liuhaotian/llava-v1.5-13b \ 5 | --question-file ./playground/data/eval/mm-vet/llava-mm-vet.jsonl \ 6 | --image-folder ./playground/data/eval/mm-vet/images \ 7 | --answers-file ./playground/data/eval/mm-vet/answers/llava-v1.5-13b.jsonl \ 8 | --temperature 0 \ 9 | --conv-mode vicuna_v1 10 | 11 | mkdir -p ./playground/data/eval/mm-vet/results 12 | 13 | python scripts/convert_mmvet_for_eval.py \ 14 | --src ./playground/data/eval/mm-vet/answers/llava-v1.5-13b.jsonl \ 15 | --dst ./playground/data/eval/mm-vet/results/llava-v1.5-13b.json 16 | 17 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/pope.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | #--model-base ../pretrained_weights/llava1.5_7b \ 3 | #--model-path ./checkpoints/dpo/llava1.5_7b-lora32-lr2e-6-lrv_5w-1e/ \ 4 | python -m llava.eval.model_vqa_loader \ 5 | --model-path ../pretrained_weights/llava1.5_7b \ 6 | --question-file ./playground/data/eval/pope/llava_pope_test.jsonl \ 7 | --image-folder ../data/coco/val2014 \ 8 | --answers-file ./playground/data/eval/pope/answers/dpo/llava1.5_7b-lora32-lr2e-6-lrv_5w-1e.jsonl \ 9 | --temperature 0 \ 10 | --conv-mode vicuna_v1 11 | 12 | python llava/eval/eval_pope.py \ 13 | --annotation-dir ./playground/data/eval/pope/coco \ 14 | --question-file ./playground/data/eval/pope/llava_pope_test.jsonl \ 15 | --result-file ./playground/data/eval/pope/answers/llava1.5_7b-lora32-lr2e-6-lrv_5w-1e.jsonl 16 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/qbench.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | if [ "$1" = "dev" ]; then 4 | echo "Evaluating in 'dev' split." 5 | elif [ "$1" = "test" ]; then 6 | echo "Evaluating in 'test' split." 7 | else 8 | echo "Unknown split, please choose between 'dev' and 'test'." 9 | exit 1 10 | fi 11 | 12 | python -m llava.eval.model_vqa_qbench \ 13 | --model-path liuhaotian/llava-v1.5-13b \ 14 | --image-folder ./playground/data/eval/qbench/images_llvisionqa/ \ 15 | --questions-file ./playground/data/eval/qbench/llvisionqa_$1.json \ 16 | --answers-file ./playground/data/eval/qbench/llvisionqa_$1_answers.jsonl \ 17 | --conv-mode llava_v1 \ 18 | --lang en 19 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/qbench_zh.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | if [ "$1" = "dev" ]; then 4 | ZH_SPLIT="验证集" 5 | echo "Evaluating in 'dev' split." 6 | elif [ "$1" = "test" ]; then 7 | ZH_SPLIT="测试集" 8 | echo "Evaluating in 'test' split." 9 | else 10 | echo "Unknown split, please choose between 'dev' and 'test'." 11 | exit 1 12 | fi 13 | 14 | python -m llava.eval.model_vqa_qbench \ 15 | --model-path liuhaotian/llava-v1.5-13b \ 16 | --image-folder ./playground/data/eval/qbench/images_llvisionqa/ \ 17 | --questions-file ./playground/data/eval/qbench/质衡-问答-$ZH_SPLIT.json \ 18 | --answers-file ./playground/data/eval/qbench/llvisionqa_zh_$1_answers.jsonl \ 19 | --conv-mode llava_v1 \ 20 | --lang zh 21 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/score_multi.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | # Check if three arguments are passed 4 | if [ "$#" -ne 7 ]; then 5 | echo "Usage: $0 " 6 | exit 1 7 | fi 8 | 9 | # Assign the command line arguments to variables 10 | model_base=$1 11 | model_path=$2 12 | question_path=$3 13 | base_answer_path=$4 14 | image_folder=$5 15 | N=$6 16 | GS=$7 17 | 18 | # Loop over each chunk/process 19 | for (( chunk_id=0; chunk_id "${base_answer_path}_merged.jsonl" 43 | for ((i=0; i> "${base_answer_path}_merged.jsonl" 46 | done 47 | # remove the unmerged files 48 | for (( chunk_id=0; chunk_id "$output_file" 28 | 29 | # Loop through the indices and concatenate each file. 30 | for IDX in $(seq 0 $((CHUNKS-1))); do 31 | cat ./playground/data/eval/seed_bench/answers/$CKPT/${CHUNKS}_${IDX}.jsonl >> "$output_file" 32 | done 33 | 34 | # Evaluate 35 | python scripts/convert_seed_for_submission.py \ 36 | --annotation-file ./playground/data/eval/seed_bench/SEED-Bench.json \ 37 | --result-file $output_file \ 38 | --result-upload-file ./playground/data/eval/seed_bench/answers_upload/llava-v1.5-13b.jsonl 39 | 40 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/sqa.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | python -m llava.eval.model_vqa_science \ 4 | --model-path liuhaotian/llava-v1.5-13b \ 5 | --question-file ./playground/data/eval/scienceqa/llava_test_CQM-A.json \ 6 | --image-folder ./playground/data/eval/scienceqa/images/test \ 7 | --answers-file ./playground/data/eval/scienceqa/answers/llava-v1.5-13b.jsonl \ 8 | --single-pred-prompt \ 9 | --temperature 0 \ 10 | --conv-mode vicuna_v1 11 | 12 | python llava/eval/eval_science_qa.py \ 13 | --base-dir ./playground/data/eval/scienceqa \ 14 | --result-file ./playground/data/eval/scienceqa/answers/llava-v1.5-13b.jsonl \ 15 | --output-file ./playground/data/eval/scienceqa/answers/llava-v1.5-13b_output.jsonl \ 16 | --output-result ./playground/data/eval/scienceqa/answers/llava-v1.5-13b_result.json 17 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/textvqa.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | python -m llava.eval.model_vqa_loader \ 4 | --model-path liuhaotian/llava-v1.5-13b \ 5 | --question-file ./playground/data/eval/textvqa/llava_textvqa_val_v051_ocr.jsonl \ 6 | --image-folder ./playground/data/eval/textvqa/train_images \ 7 | --answers-file ./playground/data/eval/textvqa/answers/llava-v1.5-13b.jsonl \ 8 | --temperature 0 \ 9 | --conv-mode vicuna_v1 10 | 11 | python -m llava.eval.eval_textvqa \ 12 | --annotation-file ./playground/data/eval/textvqa/TextVQA_0.5.1_val.json \ 13 | --result-file ./playground/data/eval/textvqa/answers/llava-v1.5-13b.jsonl 14 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/vizwiz.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | python -m llava.eval.model_vqa_loader \ 4 | --model-path liuhaotian/llava-v1.5-13b \ 5 | --question-file ./playground/data/eval/vizwiz/llava_test.jsonl \ 6 | --image-folder ./playground/data/eval/vizwiz/test \ 7 | --answers-file ./playground/data/eval/vizwiz/answers/llava-v1.5-13b.jsonl \ 8 | --temperature 0 \ 9 | --conv-mode vicuna_v1 10 | 11 | python scripts/convert_vizwiz_for_submission.py \ 12 | --annotation-file ./playground/data/eval/vizwiz/llava_test.jsonl \ 13 | --result-file ./playground/data/eval/vizwiz/answers/llava-v1.5-13b.jsonl \ 14 | --result-upload-file ./playground/data/eval/vizwiz/answers_upload/llava-v1.5-13b.json 15 | -------------------------------------------------------------------------------- /scripts/v1_5/eval/vqav2.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | gpu_list="${CUDA_VISIBLE_DEVICES:-0}" 4 | IFS=',' read -ra GPULIST <<< "$gpu_list" 5 | 6 | CHUNKS=${#GPULIST[@]} 7 | 8 | CKPT="llava-v1.5-13b" 9 | SPLIT="llava_vqav2_mscoco_test-dev2015" 10 | 11 | for IDX in $(seq 0 $((CHUNKS-1))); do 12 | CUDA_VISIBLE_DEVICES=${GPULIST[$IDX]} python -m llava.eval.model_vqa_loader \ 13 | --model-path liuhaotian/llava-v1.5-13b \ 14 | --question-file ./playground/data/eval/vqav2/$SPLIT.jsonl \ 15 | --image-folder ./playground/data/eval/vqav2/test2015 \ 16 | --answers-file ./playground/data/eval/vqav2/answers/$SPLIT/$CKPT/${CHUNKS}_${IDX}.jsonl \ 17 | --num-chunks $CHUNKS \ 18 | --chunk-idx $IDX \ 19 | --temperature 0 \ 20 | --conv-mode vicuna_v1 & 21 | done 22 | 23 | wait 24 | 25 | output_file=./playground/data/eval/vqav2/answers/$SPLIT/$CKPT/merge.jsonl 26 | 27 | # Clear out the output file if it exists. 28 | > "$output_file" 29 | 30 | # Loop through the indices and concatenate each file. 31 | for IDX in $(seq 0 $((CHUNKS-1))); do 32 | cat ./playground/data/eval/vqav2/answers/$SPLIT/$CKPT/${CHUNKS}_${IDX}.jsonl >> "$output_file" 33 | done 34 | 35 | python scripts/convert_vqav2_for_submission.py --split $SPLIT --ckpt $CKPT 36 | 37 | -------------------------------------------------------------------------------- /scripts/v1_5/finetune.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | deepspeed llava/train/train_mem.py \ 4 | --deepspeed ./scripts/zero3.json \ 5 | --model_name_or_path lmsys/vicuna-13b-v1.5 \ 6 | --version v1 \ 7 | --data_path ./playground/data/llava_v1_5_mix665k.json \ 8 | --image_folder ./playground/data \ 9 | --vision_tower openai/clip-vit-large-patch14-336 \ 10 | --pretrain_mm_mlp_adapter ./checkpoints/llava-v1.5-13b-pretrain/mm_projector.bin \ 11 | --mm_projector_type mlp2x_gelu \ 12 | --mm_vision_select_layer -2 \ 13 | --mm_use_im_start_end False \ 14 | --mm_use_im_patch_token False \ 15 | --image_aspect_ratio pad \ 16 | --group_by_modality_length True \ 17 | --bf16 True \ 18 | --output_dir ./checkpoints/llava-v1.5-13b \ 19 | --num_train_epochs 1 \ 20 | --per_device_train_batch_size 16 \ 21 | --per_device_eval_batch_size 4 \ 22 | --gradient_accumulation_steps 1 \ 23 | --evaluation_strategy "no" \ 24 | --save_strategy "steps" \ 25 | --save_steps 50000 \ 26 | --save_total_limit 1 \ 27 | --learning_rate 2e-5 \ 28 | --weight_decay 0. \ 29 | --warmup_ratio 0.03 \ 30 | --lr_scheduler_type "cosine" \ 31 | --logging_steps 1 \ 32 | --tf32 True \ 33 | --model_max_length 2048 \ 34 | --gradient_checkpointing True \ 35 | --dataloader_num_workers 4 \ 36 | --lazy_preprocess True \ 37 | --report_to wandb 38 | -------------------------------------------------------------------------------- /scripts/v1_5/finetune_lora.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | deepspeed llava/train/train_mem.py \ 4 | --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \ 5 | --deepspeed ./scripts/zero3.json \ 6 | --model_name_or_path lmsys/vicuna-13b-v1.5 \ 7 | --version v1 \ 8 | --data_path ./playground/data/llava_v1_5_mix665k.json \ 9 | --image_folder ./playground/data \ 10 | --vision_tower openai/clip-vit-large-patch14-336 \ 11 | --pretrain_mm_mlp_adapter ./checkpoints/llava-v1.5-13b-pretrain/mm_projector.bin \ 12 | --mm_projector_type mlp2x_gelu \ 13 | --mm_vision_select_layer -2 \ 14 | --mm_use_im_start_end False \ 15 | --mm_use_im_patch_token False \ 16 | --image_aspect_ratio pad \ 17 | --group_by_modality_length True \ 18 | --bf16 True \ 19 | --output_dir ./checkpoints/llava-v1.5-13b-lora \ 20 | --num_train_epochs 1 \ 21 | --per_device_train_batch_size 16 \ 22 | --per_device_eval_batch_size 4 \ 23 | --gradient_accumulation_steps 1 \ 24 | --evaluation_strategy "no" \ 25 | --save_strategy "steps" \ 26 | --save_steps 50000 \ 27 | --save_total_limit 1 \ 28 | --learning_rate 2e-4 \ 29 | --weight_decay 0. \ 30 | --warmup_ratio 0.03 \ 31 | --lr_scheduler_type "cosine" \ 32 | --logging_steps 1 \ 33 | --tf32 True \ 34 | --model_max_length 2048 \ 35 | --gradient_checkpointing True \ 36 | --dataloader_num_workers 4 \ 37 | --lazy_preprocess True \ 38 | --report_to wandb 39 | -------------------------------------------------------------------------------- /scripts/v1_5/finetune_task.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | deepspeed llava/train/train_mem.py \ 4 | --deepspeed ./scripts/zero3.json \ 5 | --model_name_or_path liuhaotian/llava-v1.5-13b \ 6 | --version v1 \ 7 | --data_path ./playground/data/llava_v1_5_mix665k.json \ 8 | --image_folder ./playground/data \ 9 | --vision_tower openai/clip-vit-large-patch14-336 \ 10 | --mm_projector_type mlp2x_gelu \ 11 | --mm_vision_select_layer -2 \ 12 | --mm_use_im_start_end False \ 13 | --mm_use_im_patch_token False \ 14 | --image_aspect_ratio pad \ 15 | --group_by_modality_length True \ 16 | --bf16 True \ 17 | --output_dir ./checkpoints/llava-v1.5-13b-task \ 18 | --num_train_epochs 1 \ 19 | --per_device_train_batch_size 16 \ 20 | --per_device_eval_batch_size 4 \ 21 | --gradient_accumulation_steps 1 \ 22 | --evaluation_strategy "no" \ 23 | --save_strategy "steps" \ 24 | --save_steps 50000 \ 25 | --save_total_limit 1 \ 26 | --learning_rate 2e-5 \ 27 | --weight_decay 0. \ 28 | --warmup_ratio 0.03 \ 29 | --lr_scheduler_type "cosine" \ 30 | --logging_steps 1 \ 31 | --tf32 True \ 32 | --model_max_length 2048 \ 33 | --gradient_checkpointing True \ 34 | --dataloader_num_workers 4 \ 35 | --lazy_preprocess True \ 36 | --report_to wandb 37 | -------------------------------------------------------------------------------- /scripts/v1_5/finetune_task_lora.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | deepspeed llava/train/train_mem.py \ 4 | --lora_enable True --lora_r 128 --lora_alpha 256 --mm_projector_lr 2e-5 \ 5 | --deepspeed ./scripts/zero3.json \ 6 | --model_name_or_path liuhaotian/llava-v1.5-13b \ 7 | --version v1 \ 8 | --data_path ./playground/data/llava_v1_5_mix665k.json \ 9 | --image_folder ./playground/data \ 10 | --vision_tower openai/clip-vit-large-patch14-336 \ 11 | --mm_projector_type mlp2x_gelu \ 12 | --mm_vision_select_layer -2 \ 13 | --mm_use_im_start_end False \ 14 | --mm_use_im_patch_token False \ 15 | --image_aspect_ratio pad \ 16 | --group_by_modality_length True \ 17 | --bf16 True \ 18 | --output_dir ./checkpoints/llava-v1.5-13b-task-lora \ 19 | --num_train_epochs 1 \ 20 | --per_device_train_batch_size 16 \ 21 | --per_device_eval_batch_size 4 \ 22 | --gradient_accumulation_steps 1 \ 23 | --evaluation_strategy "no" \ 24 | --save_strategy "steps" \ 25 | --save_steps 50000 \ 26 | --save_total_limit 1 \ 27 | --learning_rate 2e-4 \ 28 | --weight_decay 0. \ 29 | --warmup_ratio 0.03 \ 30 | --lr_scheduler_type "cosine" \ 31 | --logging_steps 1 \ 32 | --tf32 True \ 33 | --model_max_length 2048 \ 34 | --gradient_checkpointing True \ 35 | --dataloader_num_workers 4 \ 36 | --lazy_preprocess True \ 37 | --report_to wandb 38 | -------------------------------------------------------------------------------- /scripts/v1_5/pretrain.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | 3 | deepspeed llava/train/train_mem.py \ 4 | --deepspeed ./scripts/zero2.json \ 5 | --model_name_or_path lmsys/vicuna-13b-v1.5 \ 6 | --version plain \ 7 | --data_path ./playground/data/LLaVA-Pretrain/blip_laion_cc_sbu_558k.json \ 8 | --image_folder ./playground/data/LLaVA-Pretrain/images \ 9 | --vision_tower openai/clip-vit-large-patch14-336 \ 10 | --mm_projector_type mlp2x_gelu \ 11 | --tune_mm_mlp_adapter True \ 12 | --mm_vision_select_layer -2 \ 13 | --mm_use_im_start_end False \ 14 | --mm_use_im_patch_token False \ 15 | --bf16 True \ 16 | --output_dir ./checkpoints/llava-v1.5-13b-pretrain \ 17 | --num_train_epochs 1 \ 18 | --per_device_train_batch_size 32 \ 19 | --per_device_eval_batch_size 4 \ 20 | --gradient_accumulation_steps 1 \ 21 | --evaluation_strategy "no" \ 22 | --save_strategy "steps" \ 23 | --save_steps 24000 \ 24 | --save_total_limit 1 \ 25 | --learning_rate 1e-3 \ 26 | --weight_decay 0. \ 27 | --warmup_ratio 0.03 \ 28 | --lr_scheduler_type "cosine" \ 29 | --logging_steps 1 \ 30 | --tf32 True \ 31 | --model_max_length 2048 \ 32 | --gradient_checkpointing True \ 33 | --dataloader_num_workers 4 \ 34 | --lazy_preprocess True \ 35 | --report_to wandb 36 | --------------------------------------------------------------------------------