62d4db93ba6ade9bd59be8b81ba215f81ef3a2a74f84f8d0ff5d1cc1679dfec4

.gitattributes

@@ -187,3 +187,9 @@ sft/1M3/Full_remoe/logs/0626_1526_llava_v1.5_gqa_llava_model_args_5c6951/gqa.jso
 sft/1M3/Full_remoe/checkpoint-27572/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 sft/1M3/Full_remoe/checkpoint-34462/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 sft/1M3/Full_remoe/checkpoint-20679/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/pope.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/revise_Full_remoe/logs/0625_1719_llava..._pope_llava_model_args_179bff/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/mmbench_en_dev.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a79550fd31d331e4f51c173e5642111214707a274930ee0b30257a52749e5ed4
+size 14568186

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/mmerealworld_lite.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c7040f4e44e98d0d13f9c1aef1f8dec6437872877c7aeb27a8d9c9b1a2e414e
+size 1994104243

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/pope.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2da443f1a03aa045dd38d0f23abc0fb48e95ecee0b95c0fa39b890d1c9dfde90
+size 17622143

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/rank0_metric_eval_done.txt

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

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/rank1_metric_eval_done.txt

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

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/rank2_metric_eval_done.txt

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

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/rank3_metric_eval_done.txt

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

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/results.json

@@ -0,0 +1,488 @@
+{
+    "results": {
+        "ai2d": {
+            "exact_match,flexible-extract": 0.6369818652849741,
+            "exact_match_stderr,flexible-extract": 0.008654846701304468,
+            "alias": "ai2d"
+        },
+        "mmbench_en_dev": {
+            "gpt_eval_score,none": 55.2405498281787,
+            "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": 0.2850442939030745,
+            "mme_realworld_score_stderr,none": "N/A",
+            "alias": "mmerealworld_lite"
+        },
+        "mmmu_val": {
+            "mmmu_acc,none": 0.40111,
+            "mmmu_acc_stderr,none": "N/A",
+            "alias": "mmmu_val"
+        },
+        "mmstar": {
+            "coarse perception,none": 0.5149000698074137,
+            "coarse perception_stderr,none": "N/A",
+            "fine-grained perception,none": 0.25612257430606733,
+            "fine-grained perception_stderr,none": "N/A",
+            "instance reasoning,none": 0.4350976235717084,
+            "instance reasoning_stderr,none": "N/A",
+            "logical reasoning,none": 0.3274327432743274,
+            "logical reasoning_stderr,none": "N/A",
+            "math,none": 0.35330348391695626,
+            "math_stderr,none": "N/A",
+            "science & technology,none": 0.26435070136164796,
+            "science & technology_stderr,none": "N/A",
+            "alias": "mmstar"
+        },
+        "pope": {
+            "pope_accuracy,none": 0.21433333333333332,
+            "pope_accuracy_stderr,none": "N/A",
+            "pope_precision,none": 0.7677725118483413,
+            "pope_precision_stderr,none": "N/A",
+            "pope_recall,none": 0.9055338177752935,
+            "pope_recall_stderr,none": "N/A",
+            "pope_f1_score,none": 0.830982303154655,
+            "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": 0.19080000010728837,
+            "exact_match_stderr,none": 0.005413033542389061,
+            "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 0x7fa8edb12ca0>",
+            "doc_to_text": "<function ai2d_doc_to_text at 0x7fa8edb12f70>",
+            "doc_to_target": "<function ai2d_doc_to_target at 0x7fa8edb1b550>",
+            "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 0x7fa94cd6f550>",
+            "doc_to_text": "<function mmbench_doc_to_text at 0x7fa94cd6fa60>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmbench_process_results at 0x7fa94cd6ff70>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "gpt_eval_score",
+                    "aggregation": "<function mmbench_aggregate_dev_results_eval at 0x7fa94cd65dc0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "submission",
+                    "aggregation": "<function mmbench_aggregate_dev_results_submission at 0x7fa94cd6f310>",
+                    "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 0x7fa8ee2bce50>",
+            "doc_to_text": "<function mme_realworld_doc_to_text at 0x7fa8ee2c0940>",
+            "doc_to_target": "answer",
+            "process_results": "<function mme_realworld_process_results at 0x7fa8ee2c85e0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mme_realworld_score",
+                    "aggregation": "<function mme_realworld_aggregate_results at 0x7fa8ee2cb160>",
+                    "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 0x7fa8ffbaf8b0>",
+            "doc_to_text": "<function mmmu_doc_to_text at 0x7fa8ffac45e0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmmu_process_results at 0x7fa8ff9b64c0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mmmu_acc",
+                    "aggregation": "<function mmmu_aggregate_results at 0x7fa8ff985430>",
+                    "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 0x7fa8eda73310>",
+            "doc_to_text": "<function mmstar_doc_to_text at 0x7fa8eda73790>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmstar_process_results at 0x7fa8eda73ca0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "coarse perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fa8eda7a160>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "fine-grained perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fa8eda7a550>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "instance reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fa8eda7a940>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "logical reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fa8eda7ad30>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "science & technology",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fa8eda82160>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "math",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fa8eda82550>",
+                    "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 0x7fa8ee355940>",
+            "doc_to_text": "<function pope_doc_to_text at 0x7fa8ee355ee0>",
+            "doc_to_target": "answer",
+            "process_results": "<function pope_process_results at 0x7fa8ee35c4c0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "pope_accuracy",
+                    "aggregation": "<function pope_aggregate_accuracy at 0x7fa8ee35ca60>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_precision",
+                    "aggregation": "<function pope_aggregate_precision at 0x7fa8ee35f040>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_recall",
+                    "aggregation": "<function pope_aggregate_recall at 0x7fa8ee35f5e0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_f1_score",
+                    "aggregation": "<function pope_aggregate_f1_score at 0x7fa8ee35fb80>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_yes_ratio",
+                    "aggregation": "<function pope_aggregate_yes_ratio at 0x7fa8ee362160>",
+                    "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 0x7fa91a18d940>",
+            "doc_to_text": "<function textvqa_doc_to_text at 0x7fa91a18ddc0>",
+            "doc_to_target": "answer",
+            "process_results": "<function textvqa_process_results at 0x7fa919ff50d0>",
+            "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 0x7fa91a18d700>",
+                    "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/revise_Full_remoe,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/submissions/mmbench_en_dev_results.json

@@ -0,0 +1 @@
+{"overall_acc": 0.552405498281787, "category_acc": {"action_recognition": 0.8148148148148148, "attribute_comparison": 0.38636363636363635, "attribute_recognition": 0.7297297297297297, "celebrity_recognition": 0.46464646464646464, "function_reasoning": 0.7468354430379747, "future_prediction": 0.4, "identity_reasoning": 0.9555555555555556, "image_emotion": 0.6, "image_quality": 0.22641509433962265, "image_scene": 0.8365384615384616, "image_style": 0.660377358490566, "image_topic": 0.8333333333333334, "nature_relation": 0.4166666666666667, "object_localization": 0.1728395061728395, "ocr": 0.5897435897435898, "physical_property_reasoning": 0.4666666666666667, "physical_relation": 0.5416666666666666, "social_relation": 0.6046511627906976, "spatial_relationship": 0.17777777777777778, "structuralized_imagetext_understanding": 0.3974358974358974}, "l2_category_acc": {"attribute_reasoning": 0.6884422110552764, "coarse_perception": 0.6554054054054054, "finegrained_perception (cross-instance)": 0.4825174825174825, "finegrained_perception (instance-level)": 0.46757679180887374, "logic_reasoning": 0.3983050847457627, "relation_reasoning": 0.5130434782608696}}

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/submissions/mmbench_en_dev_results.xlsx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e6d183c1f91a54aac779ec9e405db197642e01de8a36f3bc0654c84698e40f7
+size 865176

sft/1M3/revise_Full_remoe/logs/0625_1531_llava..._pope_llava_model_args_179bff/textvqa_val.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:066f410eaf5beb02ca92dc3f63ce9b8cc1a90b2cafb5f28a56bd42bf1c1cbb6c
+size 13251253

sft/1M3/revise_Full_remoe/logs/0625_1552_llava...image_llava_model_args_179bff/rank1_metric_eval_done.txt

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

sft/1M3/revise_Full_remoe/logs/0625_1552_llava...image_llava_model_args_179bff/rank2_metric_eval_done.txt

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

sft/1M3/revise_Full_remoe/logs/0625_1552_llava...image_llava_model_args_179bff/rank3_metric_eval_done.txt

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

sft/1M3/revise_Full_remoe/logs/0625_1719_llava..._pope_llava_model_args_179bff/mmbench_en_dev.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0048b4b6c19e2c7b1f8242358e855dba909a31389dff285e79988ba640e1478f
+size 14568186