Upload folder using huggingface_hub (#45)

- 4f28b41f45efd90af7654281c0949c74e74589219d084208c2e0f58d93157a66 (23158690d27b8f0a82d840a017a8035abb8eb579)
- ee5e735923a55a77addf49cac82baf7b2e69599778b668889a2b3af98ccc6498 (3b9b124ff8a325340070de4676ced04dd6c9665c)

.gitattributes

@@ -136,3 +136,11 @@ sft/1M3/Full_smoe/logs/0626_0643_llava_v1.5_pope_llava_model_args_bc59f1/pope.js
 sft/1M3/Full_smoe/checkpoint-27572/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 sft/1M3/Full_smoe/checkpoint-34462/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 sft/1M3/Full_smoe/checkpoint-20679/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/pope.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0617_0149_llava...l_mme_llava_model_args_d02902/scienceqa.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0617_0304_llava...a_val_llava_model_args_d02902/infovqa_val.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/Full_smoe_share/logs/0626_1043_llava_v1.5_gqa_llava_model_args_d02902/gqa.json filter=lfs diff=lfs merge=lfs -text

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/mmbench_en_dev.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fc60bc843577c7c503dfb50c7f7a3f29b96a206d90e1594269822f3435ff2e6
+size 14568182

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/mmerealworld_lite.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1346da7ef7ec5b60710980a820da15da8f7ec493bba685a4ad11765e7a55d56
+size 1994104215

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/pope.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f13db292a45913a03d609afc38ea46ab455b5b1835560a2526a0f15efbc9fb6b
+size 17579971

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/rank3_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 3 eval done

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/results.json

@@ -0,0 +1,488 @@
+{
+    "results": {
+        "ai2d": {
+            "exact_match,flexible-extract": 0.680699481865285,
+            "exact_match_stderr,flexible-extract": 0.008390912782034982,
+            "alias": "ai2d"
+        },
+        "mmbench_en_dev": {
+            "gpt_eval_score,none": 71.30584192439862,
+            "gpt_eval_score_stderr,none": "N/A",
+            "submission,none": null,
+            "submission_stderr,none": "N/A",
+            "alias": "mmbench_en_dev"
+        },
+        "mmerealworld_lite": {
+            "mme_realworld_score,none": 0.3392391870766024,
+            "mme_realworld_score_stderr,none": "N/A",
+            "alias": "mmerealworld_lite"
+        },
+        "mmmu_val": {
+            "mmmu_acc,none": 0.44,
+            "mmmu_acc_stderr,none": "N/A",
+            "alias": "mmmu_val"
+        },
+        "mmstar": {
+            "coarse perception,none": 0.690574239304507,
+            "coarse perception_stderr,none": "N/A",
+            "fine-grained perception,none": 0.3509702775730779,
+            "fine-grained perception_stderr,none": "N/A",
+            "instance reasoning,none": 0.5402148829115121,
+            "instance reasoning_stderr,none": "N/A",
+            "logical reasoning,none": 0.3802980298029803,
+            "logical reasoning_stderr,none": "N/A",
+            "math,none": 0.4136316047402655,
+            "math_stderr,none": "N/A",
+            "science & technology,none": 0.30111656500516865,
+            "science & technology_stderr,none": "N/A",
+            "alias": "mmstar"
+        },
+        "pope": {
+            "pope_accuracy,none": 0.8244444444444444,
+            "pope_accuracy_stderr,none": "N/A",
+            "pope_precision,none": 0.9177939646201873,
+            "pope_precision_stderr,none": "N/A",
+            "pope_recall,none": 0.7931654676258992,
+            "pope_recall_stderr,none": "N/A",
+            "pope_f1_score,none": 0.8509406657018813,
+            "pope_f1_score_stderr,none": "N/A",
+            "pope_yes_ratio,none": 0.5,
+            "pope_yes_ratio_stderr,none": "N/A",
+            "alias": "pope"
+        },
+        "textvqa_val": {
+            "exact_match,none": 0.41262000023126605,
+            "exact_match_stderr,none": 0.006750530419153633,
+            "submission,none": null,
+            "submission_stderr,none": "N/A",
+            "alias": "textvqa_val"
+        }
+    },
+    "configs": {
+        "ai2d": {
+            "task": "ai2d",
+            "dataset_path": "lmms-lab/ai2d",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function ai2d_doc_to_visual at 0x7f8d7b36b1f0>",
+            "doc_to_text": "<function ai2d_doc_to_text at 0x7f8d7b36b4c0>",
+            "doc_to_target": "<function ai2d_doc_to_target at 0x7f8d7b36ba60>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 512,
+                "temperature": 0.0,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "filter_list": [
+                {
+                    "name": "flexible-extract",
+                    "filter": [
+                        {
+                            "function": "<class 'utils.MultiChoiceRegexFilter'>",
+                            "group_select": 0,
+                            "ignore_case": true,
+                            "ignore_punctuation": true,
+                            "regex_pattern": "([A-Z])\\."
+                        }
+                    ]
+                }
+            ],
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "prompt_format": "mcq",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly."
+                },
+                "gpt4v": {
+                    "prompt_format": "mcq",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAbove choices are given in {option}. {content} format.\nPlease answer with the option letter from the given choices directly."
+                },
+                "qwen_vl": {
+                    "prompt_format": "qa",
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                },
+                "xcomposer2_4khd": {
+                    "prompt_format": "mcq_xcomposer",
+                    "pre_prompt": "[UNUSED_TOKEN_146]user\nQuestion: ",
+                    "post_prompt": "[UNUSED_TOKEN_145]\n[UNUSED_TOKEN_146]assistant\nThe answer is"
+                }
+            },
+            "model_specific_target_kwargs": {
+                "default": "mcq",
+                "qwen_vl": "qa"
+            }
+        },
+        "mmbench_en_dev": {
+            "task": "mmbench_en_dev",
+            "dataset_path": "lmms-lab/MMBench",
+            "dataset_name": "en",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "dev",
+            "doc_to_visual": "<function mmbench_doc_to_visual at 0x7f8dda5adb80>",
+            "doc_to_text": "<function mmbench_doc_to_text at 0x7f8dda5bb0d0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmbench_process_results at 0x7f8dda5bb5e0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "gpt_eval_score",
+                    "aggregation": "<function mmbench_aggregate_dev_results_eval at 0x7f8dda5ad430>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "submission",
+                    "aggregation": "<function mmbench_aggregate_dev_results_submission at 0x7f8dda5ad940>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "until": [
+                    "ASSISTANT:"
+                ],
+                "max_new_tokens": 1024,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "image_aspect_ratio": "original"
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly."
+                }
+            },
+            "model_specific_generation_kwargs": {
+                "llava": {
+                    "image_aspect_ratio": "original"
+                }
+            }
+        },
+        "mmerealworld_lite": {
+            "task": "mmerealworld_lite",
+            "dataset_path": "yifanzhang114/MME-RealWorld-lite-lmms-eval",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "train",
+            "doc_to_visual": "<function mme_realworld_doc_to_visual at 0x7f8d7bb113a0>",
+            "doc_to_text": "<function mme_realworld_doc_to_text at 0x7f8d7bb11e50>",
+            "doc_to_target": "answer",
+            "process_results": "<function mme_realworld_process_results at 0x7f8d7bb16af0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mme_realworld_score",
+                    "aggregation": "<function mme_realworld_aggregate_results at 0x7f8d7bb1d670>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nSelect the best answer to the above multiple-choice question based on the image. Respond with only the letter (A, B, C, D, or E) of the correct option."
+                },
+                "gpt4v": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nSelect the best answer to the above multiple-choice question based on the image. Respond with only the letter (A, B, C, D, or E) of the correct option."
+                },
+                "xcomposer2_4khd": {
+                    "pre_prompt": "[UNUSED_TOKEN_146]user\n",
+                    "post_prompt": " Answer this question with A, B, C, or D.[UNUSED_TOKEN_145]\n[UNUSED_TOKEN_146]assistant\n"
+                }
+            }
+        },
+        "mmmu_val": {
+            "task": "mmmu_val",
+            "dataset_path": "lmms-lab/MMMU",
+            "test_split": "validation",
+            "doc_to_visual": "<function mmmu_doc_to_visual at 0x7f8d8d6cfe50>",
+            "doc_to_text": "<function mmmu_doc_to_text at 0x7f8d8d60bb80>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmmu_process_results at 0x7f8d8d427a60>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mmmu_acc",
+                    "aggregation": "<function mmmu_aggregate_results at 0x7f8d8d3349d0>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 128,
+                "until": [
+                    "\n\n"
+                ],
+                "image_aspect_ratio": "original"
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_generation_kwargs": {
+                "llava": {
+                    "image_aspect_ratio": "original"
+                }
+            }
+        },
+        "mmstar": {
+            "task": "mmstar",
+            "dataset_path": "Lin-Chen/MMStar",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "val",
+            "doc_to_visual": "<function mmstar_doc_to_visual at 0x7f8d7b2c5820>",
+            "doc_to_text": "<function mmstar_doc_to_text at 0x7f8d7b2c5ca0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmstar_process_results at 0x7f8d7b2cb1f0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "coarse perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f8d7b2cb670>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "fine-grained perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f8d7b2cba60>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "instance reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f8d7b2cbe50>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "logical reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f8d7b2d2280>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "science & technology",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f8d7b2d2670>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "math",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f8d7b2d2a60>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "until": [
+                    "\n\n"
+                ],
+                "do_sample": false
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly"
+                }
+            }
+        },
+        "pope": {
+            "task": "pope",
+            "dataset_path": "lmms-lab/POPE",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function pope_doc_to_visual at 0x7f8d7bba7e50>",
+            "doc_to_text": "<function pope_doc_to_text at 0x7f8d7bbac430>",
+            "doc_to_target": "answer",
+            "process_results": "<function pope_process_results at 0x7f8d7bbac9d0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "pope_accuracy",
+                    "aggregation": "<function pope_aggregate_accuracy at 0x7f8d7bbacf70>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_precision",
+                    "aggregation": "<function pope_aggregate_precision at 0x7f8d7bbaf550>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_recall",
+                    "aggregation": "<function pope_aggregate_recall at 0x7f8d7bbafaf0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_f1_score",
+                    "aggregation": "<function pope_aggregate_f1_score at 0x7f8d7bbb50d0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_yes_ratio",
+                    "aggregation": "<function pope_aggregate_yes_ratio at 0x7f8d7bbb5670>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 128,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ]
+        },
+        "textvqa_val": {
+            "task": "textvqa_val",
+            "dataset_path": "lmms-lab/textvqa",
+            "test_split": "validation",
+            "doc_to_visual": "<function textvqa_doc_to_visual at 0x7f8da7492ee0>",
+            "doc_to_text": "<function textvqa_doc_to_text at 0x7f8da74ae3a0>",
+            "doc_to_target": "answer",
+            "process_results": "<function textvqa_process_results at 0x7f8da74ae670>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                },
+                {
+                    "metric": "submission",
+                    "aggregation": "<function textvqa_aggregate_submissions at 0x7f8da7492ca0>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "until": [
+                    "ASSISTANT:"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question using a single word or phrase.",
+                    "ocr": false
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                }
+            }
+        }
+    },
+    "versions": {
+        "ai2d": "Yaml",
+        "mmbench_en_dev": "Yaml",
+        "mmerealworld_lite": "Yaml",
+        "mmmu_val": "Yaml",
+        "mmstar": "Yaml",
+        "pope": "Yaml",
+        "textvqa_val": "Yaml"
+    },
+    "n-shot": {
+        "ai2d": 0,
+        "mmbench_en_dev": 0,
+        "mmerealworld_lite": 0,
+        "mmmu_val": 0,
+        "mmstar": 0,
+        "pope": 0,
+        "textvqa_val": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/SMOE/1M3/Full_smoe_share,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/submissions/mmbench_en_dev_results.json

@@ -0,0 +1 @@
+{"overall_acc": 0.7130584192439863, "category_acc": {"action_recognition": 0.8888888888888888, "attribute_comparison": 0.7727272727272727, "attribute_recognition": 0.9054054054054054, "celebrity_recognition": 0.7777777777777778, "function_reasoning": 0.8734177215189873, "future_prediction": 0.55, "identity_reasoning": 0.9777777777777777, "image_emotion": 0.8, "image_quality": 0.49056603773584906, "image_scene": 0.9423076923076923, "image_style": 0.8679245283018868, "image_topic": 0.9444444444444444, "nature_relation": 0.625, "object_localization": 0.4074074074074074, "ocr": 0.6923076923076923, "physical_property_reasoning": 0.44, "physical_relation": 0.5833333333333334, "social_relation": 0.8604651162790697, "spatial_relationship": 0.2, "structuralized_imagetext_understanding": 0.5384615384615384}, "l2_category_acc": {"attribute_reasoning": 0.7336683417085427, "coarse_perception": 0.8243243243243243, "finegrained_perception (cross-instance)": 0.6363636363636364, "finegrained_perception (instance-level)": 0.6962457337883959, "logic_reasoning": 0.5423728813559322, "relation_reasoning": 0.7043478260869566}}

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/submissions/mmbench_en_dev_results.xlsx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6992a2e413cedb6220b83b925d93c26666a901c986a1f452c56f88af189f60c6
+size 865107

sft/1M3/Full_smoe_share/logs/0616_2246_llava..._pope_llava_model_args_d02902/textvqa_val.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bdba1314d6f47d7ea359cbf78d687890c9f629c0543a9f0689865685c8c773d
+size 13164500

sft/1M3/Full_smoe_share/logs/0616_2315_llava...bench_llava_model_args_d02902/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

sft/1M3/Full_smoe_share/logs/0616_2315_llava...bench_llava_model_args_d02902/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

sft/1M3/Full_smoe_share/logs/0616_2315_llava...bench_llava_model_args_d02902/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done

sft/1M3/Full_smoe_share/logs/0616_2315_llava...bench_llava_model_args_d02902/rank3_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 3 eval done

sft/1M3/Full_smoe_share/logs/0616_2315_llava...bench_llava_model_args_d02902/results.json

@@ -0,0 +1,67 @@
+{
+    "results": {
+        "ocrbench": {
+            "ocrbench_accuracy,none": 0.408,
+            "ocrbench_accuracy_stderr,none": "N/A",
+            "alias": "ocrbench"
+        }
+    },
+    "configs": {
+        "ocrbench": {
+            "task": "ocrbench",
+            "dataset_path": "echo840/OCRBench",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function ocrbench_doc_to_visual at 0x7f522fad1310>",
+            "doc_to_text": "<function ocrbench_doc_to_text at 0x7f522f3868b0>",
+            "doc_to_target": "answer",
+            "process_results": "<function ocrbench_process_results at 0x7f522f386c10>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "ocrbench_accuracy",
+                    "aggregation": "<function ocrbench_aggregate_accuracy at 0x7f522f386f70>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 128,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ]
+        }
+    },
+    "versions": {
+        "ocrbench": "Yaml"
+    },
+    "n-shot": {
+        "ocrbench": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/SMOE/1M3/Full_smoe_share,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/Full_smoe_share/logs/0616_2315_llava...bench_llava_model_args_d02902/results/ocrbench_results.txt

@@ -0,0 +1,18 @@
+######################### OCRBench #############################
+Text Recognition(Total 300): 214
+---------------- Details of Recognition Score ------------------
+Regular Text Recognition(Total 50): 45
+Irregular Text Recognition(Total 50): 46
+Artistic Text Recognition(Total 50): 44
+Handwriting Recognition(Total 50): 19
+Digit String Recognition(Total 50): 39
+Non-Semantic Text Recognition(Total 50): 21
+----------------------------------------------------------------
+Scene Text-centric VQA(Total 200): 113
+----------------------------------------------------------------
+Doc-oriented VQA(Total 200): 26
+----------------------------------------------------------------
+Key Information Extraction(Total 200): 20
+Handwritten Mathematical Expression Recognition(Total 100): 35
+--------------------- Final Score ------------------------------
+Final Score(Total 1000): 408

sft/1M3/Full_smoe_share/logs/0617_0056_llava...a_val_llava_model_args_d02902/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

sft/1M3/Full_smoe_share/logs/0617_0057_llava...a_val_llava_model_args_d02902/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

sft/1M3/Full_smoe_share/logs/0617_0057_llava...a_val_llava_model_args_d02902/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done

sft/1M3/Full_smoe_share/logs/0617_0057_llava...a_val_llava_model_args_d02902/rank3_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 3 eval done

sft/1M3/Full_smoe_share/logs/0617_0149_llava...l_mme_llava_model_args_d02902/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

sft/1M3/Full_smoe_share/logs/0617_0149_llava...l_mme_llava_model_args_d02902/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

sft/1M3/Full_smoe_share/logs/0617_0149_llava...l_mme_llava_model_args_d02902/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done

sft/1M3/Full_smoe_share/logs/0617_0149_llava...l_mme_llava_model_args_d02902/rank3_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 3 eval done

sft/1M3/Full_smoe_share/logs/0617_0149_llava...l_mme_llava_model_args_d02902/results.json

@@ -0,0 +1,293 @@
+{
+    "results": {
+        "docvqa_val": {
+            "anls,none": 0.2809871003925967,
+            "anls_stderr,none": 0.00614632941957025,
+            "alias": "docvqa_val"
+        },
+        "mme": {
+            "mme_cognition_score,none": 342.1428571428571,
+            "mme_cognition_score_stderr,none": "N/A",
+            "mme_percetion_score,none": 1284.4185674269709,
+            "mme_percetion_score_stderr,none": "N/A",
+            "alias": "mme"
+        },
+        "scienceqa_full": {
+            "exact_match,none": 0.8541067433684884,
+            "exact_match_stderr,none": 0.01102086886933664,
+            "alias": "scienceqa_full"
+        },
+        "scienceqa": {
+            "exact_match,none": 0.8632398019335062,
+            "exact_match_stderr,none": 0.005276703120644485,
+            "alias": " - scienceqa"
+        },
+        "scienceqa_img": {
+            "exact_match,none": 0.8349033217649975,
+            "exact_match_stderr,none": 0.008268790681874767,
+            "alias": " - scienceqa_img"
+        }
+    },
+    "groups": {
+        "scienceqa_full": {
+            "exact_match,none": 0.8541067433684884,
+            "exact_match_stderr,none": 0.01102086886933664,
+            "alias": "scienceqa_full"
+        }
+    },
+    "configs": {
+        "docvqa_val": {
+            "task": "docvqa_val",
+            "dataset_path": "lmms-lab/DocVQA",
+            "dataset_name": "DocVQA",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "validation",
+            "doc_to_visual": "<function docvqa_doc_to_visual at 0x7f6c6cb88310>",
+            "doc_to_text": "<function docvqa_doc_to_text at 0x7f6c6cb88670>",
+            "doc_to_target": "answers",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "anls",
+                    "aggregation": "mean",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 32,
+                "temperature": 0.0,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question using a single word or phrase."
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                }
+            }
+        },
+        "mme": {
+            "task": "mme",
+            "dataset_path": "lmms-lab/MME",
+            "dataset_kwargs": {
+                "token": false
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function mme_doc_to_visual at 0x7f6cbb4cb0d0>",
+            "doc_to_text": "<function mme_doc_to_text at 0x7f6cbb4cb820>",
+            "doc_to_target": "answer",
+            "process_results": "<function mme_process_results at 0x7f6cbb4cbdc0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mme_percetion_score",
+                    "aggregation": "<function mme_aggregate_results at 0x7f6cb9abe310>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "mme_cognition_score",
+                    "aggregation": "<function mme_aggregate_results at 0x7f6cb9abe790>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question using a single word or phrase."
+                },
+                "gpt4v": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question with Yes or No."
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                },
+                "otterhd": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                },
+                "xcomposer2_4khd": {
+                    "pre_prompt": "[UNUSED_TOKEN_146]user\n",
+                    "post_prompt": " Answer this question briefly[UNUSED_TOKEN_145]\n[UNUSED_TOKEN_146]assistant\n"
+                }
+            }
+        },
+        "scienceqa": {
+            "task": "scienceqa",
+            "dataset_path": "lmms-lab/ScienceQA",
+            "dataset_name": "ScienceQA-FULL",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function sqa_doc_to_visual at 0x7f6c85e52790>",
+            "doc_to_text": "<function sqa_doc_to_text at 0x7f6c85e529d0>",
+            "doc_to_target": "<function sqa_doc_to_target at 0x7f6c85e52dc0>",
+            "process_results": "<function sqa_process_results at 0x7f6c85e35160>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ],
+                "image_aspect_ratio": "original"
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "format": "default",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly."
+                },
+                "qwen_vl": {
+                    "format": "qwen_vl"
+                }
+            },
+            "model_specific_generation_kwargs": {
+                "llava": {
+                    "image_aspect_ratio": "original"
+                }
+            }
+        },
+        "scienceqa_img": {
+            "task": "scienceqa_img",
+            "dataset_path": "lmms-lab/ScienceQA",
+            "dataset_name": "ScienceQA-IMG",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function sqa_doc_to_visual at 0x7f6c85e35310>",
+            "doc_to_text": "<function sqa_doc_to_text at 0x7f6c85e35550>",
+            "doc_to_target": "<function sqa_doc_to_target at 0x7f6c85e35940>",
+            "process_results": "<function sqa_process_results at 0x7f6c85e35ca0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ],
+                "image_aspect_ratio": "original"
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "format": "default",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly."
+                },
+                "qwen_vl": {
+                    "format": "qwen_vl"
+                },
+                "idefics2": {
+                    "format": "default",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer:"
+                }
+            },
+            "model_specific_generation_kwargs": {
+                "llava": {
+                    "image_aspect_ratio": "original"
+                }
+            }
+        }
+    },
+    "versions": {
+        "docvqa_val": "Yaml",
+        "mme": "Yaml",
+        "scienceqa": "Yaml",
+        "scienceqa_full": "N/A",
+        "scienceqa_img": "Yaml"
+    },
+    "n-shot": {
+        "docvqa_val": 0,
+        "mme": 0,
+        "scienceqa": 0,
+        "scienceqa_full": 0,
+        "scienceqa_img": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/SMOE/1M3/Full_smoe_share,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/Full_smoe_share/logs/0617_0149_llava...l_mme_llava_model_args_d02902/scienceqa.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b883006d8f2503020559c768dbec1385f7ccbdaf5a7ec28d2347a6ef27dc64cf
+size 11242865

sft/1M3/Full_smoe_share/logs/0617_0253_llava...l_mme_llava_model_args_d02902/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

sft/1M3/Full_smoe_share/logs/0617_0253_llava...l_mme_llava_model_args_d02902/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

sft/1M3/Full_smoe_share/logs/0617_0253_llava...l_mme_llava_model_args_d02902/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done

sft/1M3/Full_smoe_share/logs/0617_0253_llava...l_mme_llava_model_args_d02902/rank3_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 3 eval done

sft/1M3/Full_smoe_share/logs/0617_0253_llava...l_mme_llava_model_args_d02902/results.json

@@ -0,0 +1,215 @@
+{
+    "results": {
+        "docvqa_val": {
+            "anls,none": 0.2809871003925967,
+            "anls_stderr,none": 0.00614632941957025,
+            "alias": "docvqa_val"
+        },
+        "mme": {
+            "mme_cognition_score,none": 342.1428571428571,
+            "mme_cognition_score_stderr,none": "N/A",
+            "mme_percetion_score,none": 1284.4185674269709,
+            "mme_percetion_score_stderr,none": "N/A",
+            "alias": "mme"
+        },
+        "scienceqa_img": {
+            "exact_match,none": 0.8349033217649975,
+            "exact_match_stderr,none": 0.008268790681874767,
+            "alias": "scienceqa_img"
+        }
+    },
+    "configs": {
+        "docvqa_val": {
+            "task": "docvqa_val",
+            "dataset_path": "lmms-lab/DocVQA",
+            "dataset_name": "DocVQA",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "validation",
+            "doc_to_visual": "<function docvqa_doc_to_visual at 0x7fa65bd3c9d0>",
+            "doc_to_text": "<function docvqa_doc_to_text at 0x7fa65bd3cd30>",
+            "doc_to_target": "answers",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "anls",
+                    "aggregation": "mean",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 32,
+                "temperature": 0.0,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question using a single word or phrase."
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                }
+            }
+        },
+        "mme": {
+            "task": "mme",
+            "dataset_path": "lmms-lab/MME",
+            "dataset_kwargs": {
+                "token": false
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function mme_doc_to_visual at 0x7fa685983e50>",
+            "doc_to_text": "<function mme_doc_to_text at 0x7fa685941700>",
+            "doc_to_target": "answer",
+            "process_results": "<function mme_process_results at 0x7fa685941ca0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mme_percetion_score",
+                    "aggregation": "<function mme_aggregate_results at 0x7fa6848f01f0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "mme_cognition_score",
+                    "aggregation": "<function mme_aggregate_results at 0x7fa6848f0670>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question using a single word or phrase."
+                },
+                "gpt4v": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question with Yes or No."
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                },
+                "otterhd": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                },
+                "xcomposer2_4khd": {
+                    "pre_prompt": "[UNUSED_TOKEN_146]user\n",
+                    "post_prompt": " Answer this question briefly[UNUSED_TOKEN_145]\n[UNUSED_TOKEN_146]assistant\n"
+                }
+            }
+        },
+        "scienceqa_img": {
+            "task": "scienceqa_img",
+            "dataset_path": "lmms-lab/ScienceQA",
+            "dataset_name": "ScienceQA-IMG",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function sqa_doc_to_visual at 0x7fa64fedf160>",
+            "doc_to_text": "<function sqa_doc_to_text at 0x7fa64fedf3a0>",
+            "doc_to_target": "<function sqa_doc_to_target at 0x7fa64fedf790>",
+            "process_results": "<function sqa_process_results at 0x7fa64fedfaf0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ],
+                "image_aspect_ratio": "original"
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "format": "default",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly."
+                },
+                "qwen_vl": {
+                    "format": "qwen_vl"
+                },
+                "idefics2": {
+                    "format": "default",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer:"
+                }
+            },
+            "model_specific_generation_kwargs": {
+                "llava": {
+                    "image_aspect_ratio": "original"
+                }
+            }
+        }
+    },
+    "versions": {
+        "docvqa_val": "Yaml",
+        "mme": "Yaml",
+        "scienceqa_img": "Yaml"
+    },
+    "n-shot": {
+        "docvqa_val": 0,
+        "mme": 0,
+        "scienceqa_img": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/SMOE/1M3/Full_smoe_share,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/Full_smoe_share/logs/0617_0304_llava...a_val_llava_model_args_d02902/infovqa_val.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec3879886871dc449b4a8d743cc3bb332066117bd7a75027f0e96c6b0c5bb1a2
+size 576438238

sft/1M3/Full_smoe_share/logs/0617_0304_llava...a_val_llava_model_args_d02902/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

sft/1M3/Full_smoe_share/logs/0617_0304_llava...a_val_llava_model_args_d02902/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

sft/1M3/Full_smoe_share/logs/0617_0304_llava...a_val_llava_model_args_d02902/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done