| { | |
| "run": { | |
| "task": "image_text_pretrain", | |
| "lr_sched": "linear_warmup_cosine_lr", | |
| "init_lr": 3e-05, | |
| "min_lr": 1e-05, | |
| "warmup_lr": 1e-06, | |
| "weight_decay": 0.05, | |
| "max_epoch": 100, | |
| "iters_per_epoch": 500, | |
| "batch_size_train": 8, | |
| "batch_size_eval": 1, | |
| "num_workers": 4, | |
| "warmup_steps": 200, | |
| "seed": 42, | |
| "output_dir": "output/xraygpt_vqa_finetune", | |
| "amp": true, | |
| "resume_ckpt_path": "24_10_30_17_10_39_epoch20.pth", | |
| "evaluate": false, | |
| "train_splits": [ | |
| "train" | |
| ], | |
| "device": "cuda", | |
| "world_size": 8, | |
| "dist_url": "env://", | |
| "distributed": true, | |
| "rank": 0, | |
| "gpu": 0, | |
| "dist_backend": "nccl" | |
| }, | |
| "model": { | |
| "arch": "mini_gpt4", | |
| "image_size": 224, | |
| "drop_path_rate": 0, | |
| "use_grad_checkpoint": false, | |
| "vit_precision": "fp16", | |
| "freeze_vit": true, | |
| "freeze_qformer": true, | |
| "num_query_token": 32, | |
| "llama_model": "Joycean0301/Llama-3.2-3B-Instruct-Medical-Conversational", | |
| "prompt": "", | |
| "lora_r": 16, | |
| "lora_alpha": 16, | |
| "lora_target_modules": [ | |
| "q_proj", | |
| "k_proj", | |
| "v_proj", | |
| "o_proj", | |
| "gate_proj", | |
| "up_proj", | |
| "down_proj" | |
| ], | |
| "lora_dropout": 0.1, | |
| "lora_bias": "none", | |
| "model_type": "pretrain_vicuna", | |
| "max_txt_len": 160, | |
| "end_sym": "<|eot_id|>", | |
| "ckpt": "24_10_29_03_29_41_epoch160.pth" | |
| }, | |
| "preprocess": { | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| }, | |
| "eval": { | |
| "name": "blip2_image_eval", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| }, | |
| "eval": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| }, | |
| "datasets": { | |
| "vqa": { | |
| "data_type": "images", | |
| "build_info": { | |
| "storage": "dataset/mimic_finetune" | |
| }, | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| } | |
| } | |
| } | |
| { | |
| "run": { | |
| "task": "image_text_pretrain", | |
| "lr_sched": "linear_warmup_cosine_lr", | |
| "init_lr": 3e-05, | |
| "min_lr": 1e-05, | |
| "warmup_lr": 1e-06, | |
| "weight_decay": 0.05, | |
| "max_epoch": 100, | |
| "iters_per_epoch": 500, | |
| "batch_size_train": 10, | |
| "batch_size_eval": 1, | |
| "num_workers": 4, | |
| "warmup_steps": 200, | |
| "seed": 42, | |
| "output_dir": "output/xraygpt_vqa_finetune", | |
| "amp": true, | |
| "resume_ckpt_path": "24_10_30_17_10_39_epoch20.pth", | |
| "evaluate": false, | |
| "train_splits": [ | |
| "train" | |
| ], | |
| "device": "cuda", | |
| "world_size": 8, | |
| "dist_url": "env://", | |
| "distributed": true, | |
| "rank": 0, | |
| "gpu": 0, | |
| "dist_backend": "nccl" | |
| }, | |
| "model": { | |
| "arch": "mini_gpt4", | |
| "image_size": 224, | |
| "drop_path_rate": 0, | |
| "use_grad_checkpoint": false, | |
| "vit_precision": "fp16", | |
| "freeze_vit": true, | |
| "freeze_qformer": true, | |
| "num_query_token": 32, | |
| "llama_model": "Joycean0301/Llama-3.2-3B-Instruct-Medical-Conversational", | |
| "prompt": "", | |
| "lora_r": 16, | |
| "lora_alpha": 16, | |
| "lora_target_modules": [ | |
| "q_proj", | |
| "k_proj", | |
| "v_proj", | |
| "o_proj", | |
| "gate_proj", | |
| "up_proj", | |
| "down_proj" | |
| ], | |
| "lora_dropout": 0.1, | |
| "lora_bias": "none", | |
| "model_type": "pretrain_vicuna", | |
| "max_txt_len": 160, | |
| "end_sym": "<|eot_id|>", | |
| "ckpt": "24_10_29_03_29_41_epoch160.pth" | |
| }, | |
| "preprocess": { | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| }, | |
| "eval": { | |
| "name": "blip2_image_eval", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| }, | |
| "eval": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| }, | |
| "datasets": { | |
| "vqa": { | |
| "data_type": "images", | |
| "build_info": { | |
| "storage": "dataset/mimic_finetune" | |
| }, | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| } | |
| } | |
| } | |
| { | |
| "run": { | |
| "task": "image_text_pretrain", | |
| "lr_sched": "linear_warmup_cosine_lr", | |
| "init_lr": 3e-05, | |
| "min_lr": 1e-05, | |
| "warmup_lr": 1e-06, | |
| "weight_decay": 0.05, | |
| "max_epoch": 100, | |
| "iters_per_epoch": 1050, | |
| "batch_size_train": 8, | |
| "batch_size_eval": 1, | |
| "num_workers": 4, | |
| "warmup_steps": 200, | |
| "seed": 42, | |
| "output_dir": "output/xraygpt_vqa_finetune", | |
| "amp": true, | |
| "resume_ckpt_path": "24_10_30_17_10_39_epoch20.pth", | |
| "evaluate": false, | |
| "train_splits": [ | |
| "train" | |
| ], | |
| "device": "cuda", | |
| "world_size": 8, | |
| "dist_url": "env://", | |
| "distributed": true, | |
| "rank": 0, | |
| "gpu": 0, | |
| "dist_backend": "nccl" | |
| }, | |
| "model": { | |
| "arch": "mini_gpt4", | |
| "image_size": 224, | |
| "drop_path_rate": 0, | |
| "use_grad_checkpoint": false, | |
| "vit_precision": "fp16", | |
| "freeze_vit": true, | |
| "freeze_qformer": true, | |
| "num_query_token": 32, | |
| "llama_model": "Joycean0301/Llama-3.2-3B-Instruct-Medical-Conversational", | |
| "prompt": "", | |
| "lora_r": 16, | |
| "lora_alpha": 16, | |
| "lora_target_modules": [ | |
| "q_proj", | |
| "k_proj", | |
| "v_proj", | |
| "o_proj", | |
| "gate_proj", | |
| "up_proj", | |
| "down_proj" | |
| ], | |
| "lora_dropout": 0.1, | |
| "lora_bias": "none", | |
| "model_type": "pretrain_vicuna", | |
| "max_txt_len": 160, | |
| "end_sym": "<|eot_id|>", | |
| "ckpt": "24_10_29_03_29_41_epoch160.pth" | |
| }, | |
| "preprocess": { | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| }, | |
| "eval": { | |
| "name": "blip2_image_eval", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| }, | |
| "eval": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| }, | |
| "datasets": { | |
| "vqa": { | |
| "data_type": "images", | |
| "build_info": { | |
| "storage": "dataset/mimic_finetune" | |
| }, | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| } | |
| } | |
| } | |
| { | |
| "run": { | |
| "task": "image_text_pretrain", | |
| "lr_sched": "linear_warmup_cosine_lr", | |
| "init_lr": 3e-05, | |
| "min_lr": 1e-05, | |
| "warmup_lr": 1e-06, | |
| "weight_decay": 0.05, | |
| "max_epoch": 40, | |
| "iters_per_epoch": 1300, | |
| "batch_size_train": 8, | |
| "batch_size_eval": 1, | |
| "num_workers": 4, | |
| "warmup_steps": 200, | |
| "seed": 42, | |
| "output_dir": "output/xraygpt_vqa_finetune", | |
| "amp": true, | |
| "resume_ckpt_path": "24_10_30_17_10_39_epoch20.pth", | |
| "evaluate": false, | |
| "train_splits": [ | |
| "train" | |
| ], | |
| "device": "cuda", | |
| "world_size": 8, | |
| "dist_url": "env://", | |
| "distributed": true, | |
| "rank": 0, | |
| "gpu": 0, | |
| "dist_backend": "nccl" | |
| }, | |
| "model": { | |
| "arch": "mini_gpt4", | |
| "image_size": 224, | |
| "drop_path_rate": 0, | |
| "use_grad_checkpoint": false, | |
| "vit_precision": "fp16", | |
| "freeze_vit": true, | |
| "freeze_qformer": true, | |
| "num_query_token": 32, | |
| "llama_model": "Joycean0301/Llama-3.2-3B-Instruct-Medical-Conversational", | |
| "prompt": "", | |
| "lora_r": 16, | |
| "lora_alpha": 16, | |
| "lora_target_modules": [ | |
| "q_proj", | |
| "k_proj", | |
| "v_proj", | |
| "o_proj", | |
| "gate_proj", | |
| "up_proj", | |
| "down_proj" | |
| ], | |
| "lora_dropout": 0.1, | |
| "lora_bias": "none", | |
| "model_type": "pretrain_vicuna", | |
| "max_txt_len": 160, | |
| "end_sym": "<|eot_id|>", | |
| "ckpt": "24_10_29_03_29_41_epoch160.pth" | |
| }, | |
| "preprocess": { | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| }, | |
| "eval": { | |
| "name": "blip2_image_eval", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| }, | |
| "eval": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| }, | |
| "datasets": { | |
| "vqa": { | |
| "data_type": "images", | |
| "build_info": { | |
| "storage": "dataset/mimic_finetune" | |
| }, | |
| "vis_processor": { | |
| "train": { | |
| "name": "blip2_image_train", | |
| "image_size": 224 | |
| } | |
| }, | |
| "text_processor": { | |
| "train": { | |
| "name": "blip_caption" | |
| } | |
| } | |
| } | |
| } | |
| } | |
| {"train_lr": "0.000019", "train_loss": "0.684137"} | |
| {"train_lr": "0.000018", "train_loss": "0.654648"} | |
| {"train_lr": "0.000017", "train_loss": "0.626471"} | |
| {"train_lr": "0.000017", "train_loss": "0.601327"} | |
| {"train_lr": "0.000016", "train_loss": "0.576456"} | |
| {"train_lr": "0.000015", "train_loss": "0.553397"} | |
| {"train_lr": "0.000014", "train_loss": "0.530701"} | |
| {"train_lr": "0.000014", "train_loss": "0.510915"} | |
| {"train_lr": "0.000013", "train_loss": "0.490313"} | |
| {"train_lr": "0.000013", "train_loss": "0.472298"} | |
| {"train_lr": "0.000012", "train_loss": "0.453687"} | |
| {"train_lr": "0.000012", "train_loss": "0.436265"} | |
| {"train_lr": "0.000011", "train_loss": "0.421915"} | |
| {"train_lr": "0.000011", "train_loss": "0.406703"} | |
| {"train_lr": "0.000011", "train_loss": "0.392106"} | |
| {"train_lr": "0.000010", "train_loss": "0.379085"} | |
| {"train_lr": "0.000010", "train_loss": "0.367431"} | |
| {"train_lr": "0.000010", "train_loss": "0.357473"} | |
| {"train_lr": "0.000010", "train_loss": "0.345610"} | |