From 8c2297750ffe6c4d9081cabc8acb54bdd712ba03 Mon Sep 17 00:00:00 2001
From: Deyao Zhu <deyao.zhu@kaust.edu.sa>
Date: Mon, 23 Oct 2023 20:42:36 +0300
Subject: [PATCH 01/18] update env name to minigptv, add bug report and feature
 request back

---
 minigpt4/models/minigpt_base.py                    |  3 ---
 train.py                                           | 14 ++++----------
 train_configs/minigpt4_llama2_stage1_pretrain.yaml |  5 ++++-
 train_configs/minigpt4_llama2_stage2_finetune.yaml |  5 ++++-
 train_configs/minigpt4_stage1_pretrain.yaml        |  5 ++++-
 train_configs/minigpt4_stage2_finetune.yaml        |  5 ++++-
 train_configs/minigpt_v2_finetune.yaml             |  6 ++++--
 7 files changed, 24 insertions(+), 19 deletions(-)

diff --git a/minigpt4/models/minigpt_base.py b/minigpt4/models/minigpt_base.py
index b5dd9d4..cd051ec 100644
--- a/minigpt4/models/minigpt_base.py
+++ b/minigpt4/models/minigpt_base.py
@@ -178,7 +178,6 @@ class MiniGPTBase(BaseModel):
             answers = [self.llama_tokenizer(a + self.end_sym,
                                             return_tensors="pt",
                                             add_special_tokens=False).to(self.device) for a in answers]
-
             cur_id = []
             cur_target = []
             for i in range(len(questions)):
@@ -226,8 +225,6 @@ class MiniGPTBase(BaseModel):
 
             conv_q = [[self.prompt_template.format(item) for item in items] for items in conv_q]
 
-        
-
             cond_embeds, cond_atts = self.prompt_wrap(img_embeds, img_atts, [q[0] for q in conv_q])
             regress_token_ids, regress_atts, part_targets = self.tokenize_conversation(conv_q, conv_a)
 
diff --git a/train.py b/train.py
index 072a078..0013079 100644
--- a/train.py
+++ b/train.py
@@ -12,6 +12,7 @@ import random
 import numpy as np
 import torch
 import torch.backends.cudnn as cudnn
+import wandb
 
 import minigpt4.tasks as tasks
 from minigpt4.common.config import Config
@@ -30,7 +31,6 @@ from minigpt4.models import *
 from minigpt4.processors import *
 from minigpt4.runners import *
 from minigpt4.tasks import *
-import wandb
 
 
 def parse_args():
@@ -44,12 +44,10 @@ def parse_args():
         "in xxx=yyy format will be merged into config file (deprecate), "
         "change to --cfg-options instead.",
     )
-    parser.add_argument("--wandb_log", default=False)
-    parser.add_argument("--job_name",default="minigpt_v2",type=str)
+    parser.add_argument("--job_name", default="minigpt_v2",type=str)
 
     args = parser.parse_args()
 
-
     return args
 
 
@@ -80,16 +78,13 @@ def main():
     # set before init_distributed_mode() to ensure the same job_id shared across all ranks.
     job_id = now()
     args = parse_args()
-
-    cfg = Config(parse_args())
+    cfg = Config(args)
 
     init_distributed_mode(cfg.run_cfg)
-
     setup_seeds(cfg)
 
     # set after init_distributed_mode() to only log on master.
     setup_logger()
-
     cfg.pretty_print()
 
     task = tasks.setup_task(cfg)
@@ -98,10 +93,9 @@ def main():
 
     if cfg.run_cfg.wandb_log:
         wandb.login()
-        wandb.init(project="minigptv2",name=args.job_name)
+        wandb.init(project="minigptv", name=cfg.run_cfg.job_name)
         wandb.watch(model)
 
-
     runner = get_runner_class(cfg)(
         cfg=cfg, job_id=job_id, task=task, model=model, datasets=datasets
     )
diff --git a/train_configs/minigpt4_llama2_stage1_pretrain.yaml b/train_configs/minigpt4_llama2_stage1_pretrain.yaml
index c13d31f..bcc458e 100644
--- a/train_configs/minigpt4_llama2_stage1_pretrain.yaml
+++ b/train_configs/minigpt4_llama2_stage1_pretrain.yaml
@@ -52,4 +52,7 @@ run:
   device: "cuda"
   world_size: 1
   dist_url: "env://"
-  distributed: True
\ No newline at end of file
+  distributed: True
+
+  wandb_log: True
+  job_name: minigpt4_llama2_pretrain
\ No newline at end of file
diff --git a/train_configs/minigpt4_llama2_stage2_finetune.yaml b/train_configs/minigpt4_llama2_stage2_finetune.yaml
index 8c138ae..29b5358 100644
--- a/train_configs/minigpt4_llama2_stage2_finetune.yaml
+++ b/train_configs/minigpt4_llama2_stage2_finetune.yaml
@@ -46,4 +46,7 @@ run:
   device: "cuda"
   world_size: 1
   dist_url: "env://"
-  distributed: True
\ No newline at end of file
+  distributed: True
+
+  wandb_log: True
+  job_name: minigpt4_llama2_finetune
\ No newline at end of file
diff --git a/train_configs/minigpt4_stage1_pretrain.yaml b/train_configs/minigpt4_stage1_pretrain.yaml
index ce8bc87..bd9a451 100644
--- a/train_configs/minigpt4_stage1_pretrain.yaml
+++ b/train_configs/minigpt4_stage1_pretrain.yaml
@@ -52,4 +52,7 @@ run:
   device: "cuda"
   world_size: 1
   dist_url: "env://"
-  distributed: True
\ No newline at end of file
+  distributed: True
+
+  wandb_log: True
+  job_name: minigpt4_pretrain
\ No newline at end of file
diff --git a/train_configs/minigpt4_stage2_finetune.yaml b/train_configs/minigpt4_stage2_finetune.yaml
index 531a3a0..89d1100 100644
--- a/train_configs/minigpt4_stage2_finetune.yaml
+++ b/train_configs/minigpt4_stage2_finetune.yaml
@@ -46,4 +46,7 @@ run:
   device: "cuda"
   world_size: 1
   dist_url: "env://"
-  distributed: True
\ No newline at end of file
+  distributed: True
+
+  wandb_log: True
+  job_name: minigpt4_finetune
\ No newline at end of file
diff --git a/train_configs/minigpt_v2_finetune.yaml b/train_configs/minigpt_v2_finetune.yaml
index 7bf6bbf..4039ea6 100644
--- a/train_configs/minigpt_v2_finetune.yaml
+++ b/train_configs/minigpt_v2_finetune.yaml
@@ -276,7 +276,6 @@ run:
   init_lr: 1e-5
   min_lr: 8e-5
   warmup_lr: 1e-6
-  wandb_log: True
 
   weight_decay: 0.05
   max_epoch: 50
@@ -296,4 +295,7 @@ run:
   device: "cuda"
   world_size: 1
   dist_url: "env://"
-  distributed: True
\ No newline at end of file
+  distributed: True
+
+  wandb_log: True
+  job_name: minigptv2_finetune
\ No newline at end of file

From f0b6a9e7d77747ad1d73c02c35f04b41a6c23be4 Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:11:35 +0300
Subject: [PATCH 02/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 29 ++++++++++++++++++++++++----
 1 file changed, 25 insertions(+), 4 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 2d5c825..2bae2d7 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -1,23 +1,44 @@
 ## Download the COCO captions, RefCOCO, RefCOCO+. RefCOCOg, visual genome, textcaps, LLaVA, gqa, AOK-VQA, OK-VQA, OCR-VQA, filtered Flickr-30k, multi-task conversation, and Unnatural instruction datasets
 
-### COCO captions
+After downloading all of them, organize the data as follows in `./playground/data`,
 
+```
+├── coco
+│   └── train2017
+├── gqa
+│   └── images
+├── ocr_vqa
+│   └── images
+├── textvqa
+│   └── train_images
+└── vg
+    ├── VG_100K
+    └── VG_100K_2
+
+
+### COCO captions
+- [train2017](http://images.cocodataset.org/zips/train2017.zip)
 
 ### RefCOCO, RefCOCO+, RefCOCOg
 
 ### Visual genome
-
-### textcaps
+- [part1](https://cs.stanford.edu/people/rak248/VG_100K_2/images.zip), [part2](https://cs.stanford.edu/people/rak248/VG_100K_2/images2.zip)
+### TextCaps
 
 ### LLaVA
 
-### gqa
+### TextVQA
+- [train_val_images](https://dl.fbaipublicfiles.com/textvqa/images/train_val_images.zip)
+### GQA
+- [images](https://downloads.cs.stanford.edu/nlp/data/gqa/images.zip)
+- [Annotations](https://storage.googleapis.com/sfr-vision-language-research/LAVIS/datasets/gqa/testdev_balanced_questions.json)
 
 ### OKVQA
 
 ### AOK-VQA
 
 ### OCR-VQA
+- [download script](https://drive.google.com/drive/folders/1_GYPY5UkUy7HIcR0zq3ZCFgeZN7BAfm_?usp=sharing), **we save all files as `.jpg`**
 
 ### filtered Flickr-30k
 

From 68df270f14dfd30213f95349c458bd1eced6c602 Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:15:14 +0300
Subject: [PATCH 03/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 2bae2d7..438240f 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -14,7 +14,7 @@ After downloading all of them, organize the data as follows in `./playground/dat
 └── vg
     ├── VG_100K
     └── VG_100K_2
-
+```
 
 ### COCO captions
 - [train2017](http://images.cocodataset.org/zips/train2017.zip)

From c2de397a8674b1546fad0f6f01088a42efde98cf Mon Sep 17 00:00:00 2001
From: Deyao Zhu <deyao.zhu@kaust.edu.sa>
Date: Mon, 23 Oct 2023 21:31:13 +0300
Subject: [PATCH 04/18] update refcoco preparation

---
 dataset/README_MINIGPTv2_FINETUNE.md | 46 ++++++++++++++++++++++++++--
 1 file changed, 43 insertions(+), 3 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 2d5c825..be1e09b 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -2,16 +2,56 @@
 
 ### COCO captions
 
-
 ### RefCOCO, RefCOCO+, RefCOCOg
 
-### Visual genome
+Makesure you have the COCO 2014 images first. 
+
+Then,
+download RefCOCO, RefCOCO+, and RefCOCOg annotation files in the following links.
+
+- https://bvisionweb1.cs.unc.edu/licheng/referit/data/refcoco.zip
+- https://bvisionweb1.cs.unc.edu/licheng/referit/data/refcoco+.zip
+- https://bvisionweb1.cs.unc.edu/licheng/referit/data/refcocog.zip
+
+Unzip these files to the location you like. It should have the structure like the following
+
+```
+Location_you_like
+├── refcoco
+│   ├── instances.json
+│   ├── refs(google).p
+│   └── refs(unc).p
+├── refcoco+
+│   ├── instances.json
+│   └── refs(unc).p
+└── refcocog
+    ├── instances.json
+    ├── refs(google).p
+    └── refs(umd).p
+```
+
+Set **image_path** in all the following dataset configuration files to the COCO 2014 image folder.
+Similarly, set **ann_path** in all the following configs to the above folder (Location_you_like) that contains refcoco, refcoco+, and refcocog.
+
+- [minigpt4/configs/refcoco.yaml](../minigpt4/configs/refcoco.yaml)
+- [minigpt4/configs/refcocog.yaml](../minigpt4/configs/refcocog.yaml) 
+- [minigpt4/configs/refcocop.yaml](../minigpt4/configs/refcocop.yaml)
+- [minigpt4/configs/invrefcoco.yaml](../minigpt4/configs/invrefcoco.yaml)
+- [minigpt4/configs/invrefcocog.yaml](../minigpt4/configs/invrefcocog.yaml) 
+- [minigpt4/configs/invrefcocop.yaml](../minigpt4/configs/invrefcocop.yaml)
+
+
+
+### Visual Genome
 
 ### textcaps
 
 ### LLaVA
 
-### gqa
+
+
+
+### GQA
 
 ### OKVQA
 

From 114852b529ffb2b8fd5dc68e5653fb7f6b9bdacd Mon Sep 17 00:00:00 2001
From: Deyao Zhu <deyao.zhu@kaust.edu.sa>
Date: Mon, 23 Oct 2023 21:34:47 +0300
Subject: [PATCH 05/18] update refococo

---
 dataset/README_MINIGPTv2_FINETUNE.md | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index c582acc..3d09ffe 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -19,11 +19,13 @@ After downloading all of them, organize the data as follows in `./playground/dat
 ### COCO captions
 - [train2017](http://images.cocodataset.org/zips/train2017.zip)
 
-### RefCOCO, RefCOCO+, RefCOCOg
+
 
 ### Visual genome
 - [part1](https://cs.stanford.edu/people/rak248/VG_100K_2/images.zip), [part2](https://cs.stanford.edu/people/rak248/VG_100K_2/images2.zip)
 ### TextCaps
+
+### RefCOCO, RefCOCO+, RefCOCOg
 Makesure you have the COCO 2014 images first. 
 
 Then,

From ab520c89fc76d14b623aeea0a8e3bd5134ed36fd Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:41:39 +0300
Subject: [PATCH 06/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 3d09ffe..f658adc 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -26,7 +26,7 @@ After downloading all of them, organize the data as follows in `./playground/dat
 ### TextCaps
 
 ### RefCOCO, RefCOCO+, RefCOCOg
-Makesure you have the COCO 2014 images first. 
+Make sure you have the COCO 2014 images first. 
 
 Then,
 download RefCOCO, RefCOCO+, and RefCOCOg annotation files in the following links.
@@ -71,14 +71,10 @@ Similarly, set **ann_path** in all the following configs to the above folder (Lo
 ### LLaVA
 
 ### TextVQA
-- [train_val_images](https://dl.fbaipublicfiles.com/textvqa/images/train_val_images.zip)
-### GQA
-- [images](https://downloads.cs.stanford.edu/nlp/data/gqa/images.zip)
-- [Annotations](https://storage.googleapis.com/sfr-vision-language-research/LAVIS/datasets/gqa/testdev_balanced_questions.json)
-
-
+Images, and question-answer pairs will be loaded during evaluation.
 
 ### GQA
+Images, and question-answer pairs will be loaded during evaluation.
 
 ### OKVQA
 

From 50df66e81e437091b7a55e68c793dee2b3b4f5f0 Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:42:58 +0300
Subject: [PATCH 07/18] Create Evaluation.md

---
 dataset/Evaluation.md | 1 +
 1 file changed, 1 insertion(+)
 create mode 100644 dataset/Evaluation.md

diff --git a/dataset/Evaluation.md b/dataset/Evaluation.md
new file mode 100644
index 0000000..8b13789
--- /dev/null
+++ b/dataset/Evaluation.md
@@ -0,0 +1 @@
+

From fa19bc09f21c220c564148c9d7ed86d3c1ced523 Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:45:22 +0300
Subject: [PATCH 08/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 22 ----------------------
 1 file changed, 22 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index f658adc..280622e 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -1,26 +1,10 @@
 ## Download the COCO captions, RefCOCO, RefCOCO+. RefCOCOg, visual genome, textcaps, LLaVA, gqa, AOK-VQA, OK-VQA, OCR-VQA, filtered Flickr-30k, multi-task conversation, and Unnatural instruction datasets
 
-After downloading all of them, organize the data as follows in `./playground/data`,
-
-```
-├── coco
-│   └── train2017
-├── gqa
-│   └── images
-├── ocr_vqa
-│   └── images
-├── textvqa
-│   └── train_images
-└── vg
-    ├── VG_100K
-    └── VG_100K_2
-```
 
 ### COCO captions
 - [train2017](http://images.cocodataset.org/zips/train2017.zip)
 
 
-
 ### Visual genome
 - [part1](https://cs.stanford.edu/people/rak248/VG_100K_2/images.zip), [part2](https://cs.stanford.edu/people/rak248/VG_100K_2/images2.zip)
 ### TextCaps
@@ -70,12 +54,6 @@ Similarly, set **ann_path** in all the following configs to the above folder (Lo
 
 ### LLaVA
 
-### TextVQA
-Images, and question-answer pairs will be loaded during evaluation.
-
-### GQA
-Images, and question-answer pairs will be loaded during evaluation.
-
 ### OKVQA
 
 ### AOK-VQA

From b15fec91a5f6c164aa8148b577f8b19a08217057 Mon Sep 17 00:00:00 2001
From: Deyao Zhu <deyao.zhu@kaust.edu.sa>
Date: Mon, 23 Oct 2023 21:47:29 +0300
Subject: [PATCH 09/18] update llava

---
 dataset/README_MINIGPTv2_FINETUNE.md | 30 ++++++++++++++++++++++------
 1 file changed, 24 insertions(+), 6 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 3d09ffe..512b1ff 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -23,6 +23,7 @@ After downloading all of them, organize the data as follows in `./playground/dat
 
 ### Visual genome
 - [part1](https://cs.stanford.edu/people/rak248/VG_100K_2/images.zip), [part2](https://cs.stanford.edu/people/rak248/VG_100K_2/images2.zip)
+
 ### TextCaps
 
 ### RefCOCO, RefCOCO+, RefCOCOg
@@ -55,12 +56,12 @@ Location_you_like
 Set **image_path** in all the following dataset configuration files to the COCO 2014 image folder.
 Similarly, set **ann_path** in all the following configs to the above folder (Location_you_like) that contains refcoco, refcoco+, and refcocog.
 
-- [minigpt4/configs/refcoco.yaml](../minigpt4/configs/refcoco.yaml)
-- [minigpt4/configs/refcocog.yaml](../minigpt4/configs/refcocog.yaml) 
-- [minigpt4/configs/refcocop.yaml](../minigpt4/configs/refcocop.yaml)
-- [minigpt4/configs/invrefcoco.yaml](../minigpt4/configs/invrefcoco.yaml)
-- [minigpt4/configs/invrefcocog.yaml](../minigpt4/configs/invrefcocog.yaml) 
-- [minigpt4/configs/invrefcocop.yaml](../minigpt4/configs/invrefcocop.yaml)
+- [minigpt4/configs/datasets/coco_bbox/refcoco.yaml](../minigpt4/configs/datasets/coco_bbox/refcoco.yaml)
+- [minigpt4/configs/datasets/coco_bbox/refcocog.yaml](../minigpt4/configs/datasets/coco_bbox/refcocog.yaml) 
+- [minigpt4/configs/datasets/coco_bbox/refcocop.yaml](../minigpt4/configs/datasets/coco_bbox/refcocop.yaml)
+- [minigpt4/configs/datasets/coco_bbox/invrefcoco.yaml](../minigpt4/configs/datasets/coco_bbox/invrefcoco.yaml)
+- [minigpt4/configs/datasets/coco_bbox/invrefcocog.yaml](../minigpt4/configs/datasets/coco_bbox/invrefcocog.yaml) 
+- [minigpt4/configs/datasets/coco_bbox/invrefcocop.yaml](../minigpt4/configs/datasets/coco_bbox/invrefcocop.yaml)
 
 
 
@@ -69,6 +70,23 @@ Similarly, set **ann_path** in all the following configs to the above folder (Lo
 ### textcaps
 
 ### LLaVA
+Makesure you have the COCO 2014 images first. 
+
+Download Llava annotation files in the following link to the place you like.
+
+- https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/resolve/main/conversation_58k.json
+- https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/resolve/main/detail_23k.json
+- https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/resolve/main/complex_reasoning_77k.json
+
+Set **image_path** in all the following dataset configuration files to the COCO 2014 image folder.
+Similarly, set **ann_path** to the location of the previous downloaded conversation_58k.json, 
+detail_23k.json, and complex_reasoning_77k.json in conversation.yaml, detail.yaml, and reason.yaml, respectively.
+
+
+- [minigpt4/configs/datasets/llava/conversation.yaml](../minigpt4/configs/datasets/llava/conversation.yaml)
+- [minigpt4/configs/datasets/llava/detail.yaml](../minigpt4/configs/datasets/llava/detail.yaml) 
+- [minigpt4/configs/datasets/llava/reason.yaml](../minigpt4/configs/datasets/llava/reason.yaml)
+
 
 ### TextVQA
 - [train_val_images](https://dl.fbaipublicfiles.com/textvqa/images/train_val_images.zip)

From f13ad74b267b53f9abab23a4c48fb46796cf61d2 Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:48:58 +0300
Subject: [PATCH 10/18] Update Evaluation.md

---
 dataset/Evaluation.md | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/dataset/Evaluation.md b/dataset/Evaluation.md
index 8b13789..9e3ff86 100644
--- a/dataset/Evaluation.md
+++ b/dataset/Evaluation.md
@@ -1 +1,21 @@
 
+### OKVQA 
+
+### GQA
+Images and question-answer pairs will be loaded during the evaluation.
+'''
+python run_eval.py xxxx
+'''
+### VSR
+Images and question-answer pairs will be loaded during the evaluation.
+
+### IconVQA 
+
+### VizWiz 
+
+### HM
+
+
+
+
+

From 75f87692971a1a4b5107c1f2b8e1926b275bde2d Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:49:30 +0300
Subject: [PATCH 11/18] Update Evaluation.md

---
 dataset/Evaluation.md | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/dataset/Evaluation.md b/dataset/Evaluation.md
index 9e3ff86..e2b366c 100644
--- a/dataset/Evaluation.md
+++ b/dataset/Evaluation.md
@@ -3,11 +3,15 @@
 
 ### GQA
 Images and question-answer pairs will be loaded during the evaluation.
-'''
+```
 python run_eval.py xxxx
-'''
+```
+
 ### VSR
 Images and question-answer pairs will be loaded during the evaluation.
+```
+python run_eval.py xxxx
+```
 
 ### IconVQA 
 

From 45a97de8cc02e1b85bb92bf5a194a2569a0a5ada Mon Sep 17 00:00:00 2001
From: Deyao Zhu <deyao.zhu@kaust.edu.sa>
Date: Mon, 23 Oct 2023 21:49:33 +0300
Subject: [PATCH 12/18] remove unused pretrain set in the finetune readme

---
 dataset/README_MINIGPTv2_FINETUNE.md | 107 ---------------------------
 1 file changed, 107 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 070fbf5..924ecca 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -84,110 +84,3 @@ detail_23k.json, and complex_reasoning_77k.json in conversation.yaml, detail.yam
 ### Multi-task conversation
 
 ### Unnatural instruction
-
-
-
-
-
-
-
-
-
-
-
-
-
-### Pre-training datasets download:
-We use the filtered synthetic captions prepared by BLIP. For more details about the dataset, please refer to [BLIP](https://github.com/salesforce/BLIP).
-
-It requires ~2.3T to store LAION and CC3M+CC12M+SBU datasets
-
-Image source | Filtered synthetic caption by ViT-L
---- | :---:
-CC3M+CC12M+SBU | <a href="https://storage.googleapis.com/sfr-vision-language-research/BLIP/datasets/ccs_synthetic_filtered_large.json">Download</a>
-LAION115M |  <a href="https://storage.googleapis.com/sfr-vision-language-research/BLIP/datasets/laion_synthetic_filtered_large.json">Download</a>
-
-This will download two json files 
-```
-ccs_synthetic_filtered_large.json
-laion_synthetic_filtered_large.json
-```
-
-## prepare the data step-by-step
-
-
-### setup the dataset folder and move the annotation file to the data storage folder
-```
-export MINIGPT4_DATASET=/YOUR/PATH/FOR/LARGE/DATASET/
-mkdir ${MINIGPT4_DATASET}/cc_sbu
-mkdir ${MINIGPT4_DATASET}/laion
-mv ccs_synthetic_filtered_large.json ${MINIGPT4_DATASET}/cc_sbu
-mv laion_synthetic_filtered_large.json ${MINIGPT4_DATASET}/laion
-```
-
-### Convert the scripts to data storate folder
-```
-cp convert_cc_sbu.py ${MINIGPT4_DATASET}/cc_sbu
-cp download_cc_sbu.sh ${MINIGPT4_DATASET}/cc_sbu
-cp convert_laion.py ${MINIGPT4_DATASET}/laion
-cp download_laion.sh ${MINIGPT4_DATASET}/laion
-```
-
-
-### Convert the laion and cc_sbu annotation file format to be img2dataset format
-```
-cd ${MINIGPT4_DATASET}/cc_sbu
-python convert_cc_sbu.py
-
-cd ${MINIGPT4_DATASET}/laion
-python convert_laion.py
-```
-
-### Download the datasets with img2dataset
-```
-cd ${MINIGPT4_DATASET}/cc_sbu
-sh download_cc_sbu.sh
-cd ${MINIGPT4_DATASET}/laion
-sh download_laion.sh
-```
-
-
-The final dataset structure
-
-```
-.
-├── ${MINIGPT4_DATASET}
-│   ├── cc_sbu
-│       ├── convert_cc_sbu.py
-│       ├── download_cc_sbu.sh
-│       ├── ccs_synthetic_filtered_large.json
-│       ├── ccs_synthetic_filtered_large.tsv
-│       └── cc_sbu_dataset
-│           ├── 00000.tar
-│           ├── 00000.parquet
-│           ...
-│   ├── laion
-│       ├── convert_laion.py
-│       ├── download_laion.sh
-│       ├── laion_synthetic_filtered_large.json
-│       ├── laion_synthetic_filtered_large.tsv
-│       └── laion_dataset
-│           ├── 00000.tar
-│           ├── 00000.parquet
-│           ...
-...   
-```
-
-
-## Set up the dataset configuration files
-
-Then, set up the LAION dataset loading path in 
-[here](../minigpt4/configs/datasets/laion/defaults.yaml#L5) at Line 5 as 
-${MINIGPT4_DATASET}/laion/laion_dataset/{00000..10488}.tar
-
-and the Conceptual Captoin and SBU datasets loading path in 
-[here](../minigpt4/configs/datasets/cc_sbu/defaults.yaml#L5) at Line 5 as 
-${MINIGPT4_DATASET}/cc_sbu/cc_sbu_dataset/{00000..01255}.tar
-
-
-

From 0c27a75fd693df2eef6af871fcfe61ff6d637e87 Mon Sep 17 00:00:00 2001
From: Xiang Li <44761952+lx709@users.noreply.github.com>
Date: Mon, 23 Oct 2023 21:51:36 +0300
Subject: [PATCH 13/18] Update Evaluation.md

---
 dataset/Evaluation.md | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/dataset/Evaluation.md b/dataset/Evaluation.md
index e2b366c..34118f3 100644
--- a/dataset/Evaluation.md
+++ b/dataset/Evaluation.md
@@ -3,19 +3,20 @@
 
 ### GQA
 Images and question-answer pairs will be loaded during the evaluation.
-```
-python run_eval.py xxxx
-```
+
+``` python run_eval.py xxxx ``` 
 
 ### VSR
 Images and question-answer pairs will be loaded during the evaluation.
-```
-python run_eval.py xxxx
-```
+
+``` python run_eval.py xxxx ``` 
 
 ### IconVQA 
 
 ### VizWiz 
+1. Download [`test.json`](https://vizwiz.cs.colorado.edu/VizWiz_final/vqa_data/Annotations.zip) and extract [`test.zip`](https://vizwiz.cs.colorado.edu/VizWiz_final/images/test.zip) to `test`. Put them under `your_path/vizwiz`.
+2. Single-GPU inference.
+``` python run_eval.py xxxx ``` 
 
 ### HM
 

From 41c050de7661763df95aca8d7e6601481e33d058 Mon Sep 17 00:00:00 2001
From: ZhuDeyao <tsu.tikgiau@gmail.com>
Date: Mon, 23 Oct 2023 21:57:25 +0300
Subject: [PATCH 14/18] Update modeling_llama.py for transformers package
 compatibility

---
 minigpt4/models/modeling_llama.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/minigpt4/models/modeling_llama.py b/minigpt4/models/modeling_llama.py
index 6d28020..5d59a53 100644
--- a/minigpt4/models/modeling_llama.py
+++ b/minigpt4/models/modeling_llama.py
@@ -75,7 +75,7 @@ class LlamaForCausalLM(LlamaForCausalLMOrig):
         )
 
         hidden_states = outputs[0]
-        if self.config.pretraining_tp > 1:
+        if hasattr(self.config, 'pretraining_tp') and self.config.pretraining_tp > 1:
             lm_head_slices = self.lm_head.weight.split(self.vocab_size // self.config.pretraining_tp, dim=0)
             logits = [F.linear(hidden_states, lm_head_slices[i]) for i in range(self.config.pretraining_tp)]
             logits = torch.cat(logits, dim=-1)

From f955e62227b140135f2daa4c89acec3517c4931a Mon Sep 17 00:00:00 2001
From: XiaoqianShen <64844805+xiaoqian-shen@users.noreply.github.com>
Date: Mon, 23 Oct 2023 19:22:16 +0000
Subject: [PATCH 15/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 924ecca..9eb3dcf 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -74,8 +74,18 @@ detail_23k.json, and complex_reasoning_77k.json in conversation.yaml, detail.yam
 
 ### OKVQA
 
+- [OK-VQA Input Questions](https://okvqa.allenai.org/static/data/OpenEnded_mscoco_train2014_questions.json.zip)
+- [OK-VQA Annotations](https://okvqa.allenai.org/static/data/mscoco_train2014_annotations.json.zip)
+- Images are from COCO
+
 ### AOK-VQA
 
+```
+export AOKVQA_DIR=YOUR_DATASET_PATH
+mkdir -p ${AOKVQA_DIR}
+curl -fsSL https://prior-datasets.s3.us-east-2.amazonaws.com/aokvqa/aokvqa_v1p0.tar.gz | tar xvz -C ${AOKVQA_DIR}
+```
+
 ### OCR-VQA
 - [download script](https://drive.google.com/drive/folders/1_GYPY5UkUy7HIcR0zq3ZCFgeZN7BAfm_?usp=sharing), **we save all files as `.jpg`**
 

From ac626b3d9a4eaf883c00919d4fe654083a844af7 Mon Sep 17 00:00:00 2001
From: XiaoqianShen <64844805+xiaoqian-shen@users.noreply.github.com>
Date: Mon, 23 Oct 2023 19:27:01 +0000
Subject: [PATCH 16/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 9eb3dcf..edbdf2f 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -76,6 +76,7 @@ detail_23k.json, and complex_reasoning_77k.json in conversation.yaml, detail.yam
 
 - [OK-VQA Input Questions](https://okvqa.allenai.org/static/data/OpenEnded_mscoco_train2014_questions.json.zip)
 - [OK-VQA Annotations](https://okvqa.allenai.org/static/data/mscoco_train2014_annotations.json.zip)
+- [okvqa_train](https://storage.googleapis.com/sfr-vision-language-research/LAVIS/datasets/okvqa/okvqa_train.json)
 - Images are from COCO
 
 ### AOK-VQA

From dc84c5e5c7c95f9e084f6fed12c3b2db56ee1fcd Mon Sep 17 00:00:00 2001
From: XiaoqianShen <64844805+xiaoqian-shen@users.noreply.github.com>
Date: Mon, 23 Oct 2023 19:29:37 +0000
Subject: [PATCH 17/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index edbdf2f..0181f7d 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -10,6 +10,9 @@
 
 ### TextCaps
 
+-[TextCaps_0.1_train](https://dl.fbaipublicfiles.com/textvqa/data/textcaps/TextCaps_0.1_train.json)
+-[Images](https://dl.fbaipublicfiles.com/textvqa/images/train_val_images.zip)
+
 ### RefCOCO, RefCOCO+, RefCOCOg
 Make sure you have the COCO 2014 images first. 
 
@@ -47,12 +50,6 @@ Similarly, set **ann_path** in all the following configs to the above folder (Lo
 - [minigpt4/configs/datasets/coco_bbox/invrefcocog.yaml](../minigpt4/configs/datasets/coco_bbox/invrefcocog.yaml) 
 - [minigpt4/configs/datasets/coco_bbox/invrefcocop.yaml](../minigpt4/configs/datasets/coco_bbox/invrefcocop.yaml)
 
-
-
-### Visual Genome
-
-### textcaps
-
 ### LLaVA
 Makesure you have the COCO 2014 images first. 
 

From b93d40f23cfc69ba1c68d8fc78014099869e47f2 Mon Sep 17 00:00:00 2001
From: XiaoqianShen <64844805+xiaoqian-shen@users.noreply.github.com>
Date: Mon, 23 Oct 2023 19:31:05 +0000
Subject: [PATCH 18/18] Update README_MINIGPTv2_FINETUNE.md

---
 dataset/README_MINIGPTv2_FINETUNE.md | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/dataset/README_MINIGPTv2_FINETUNE.md b/dataset/README_MINIGPTv2_FINETUNE.md
index 0181f7d..f7af60d 100644
--- a/dataset/README_MINIGPTv2_FINETUNE.md
+++ b/dataset/README_MINIGPTv2_FINETUNE.md
@@ -10,8 +10,8 @@
 
 ### TextCaps
 
--[TextCaps_0.1_train](https://dl.fbaipublicfiles.com/textvqa/data/textcaps/TextCaps_0.1_train.json)
--[Images](https://dl.fbaipublicfiles.com/textvqa/images/train_val_images.zip)
+- [TextCaps_0.1_train](https://dl.fbaipublicfiles.com/textvqa/data/textcaps/TextCaps_0.1_train.json)
+- [Images](https://dl.fbaipublicfiles.com/textvqa/images/train_val_images.zip)
 
 ### RefCOCO, RefCOCO+, RefCOCOg
 Make sure you have the COCO 2014 images first.