Upload folder using huggingface_hub

.gitattributes

@@ -40,3 +40,9 @@ sft/smoe_refine_665k_llava/checkpoint-16632/logs/0409_1833_llava...mstar_llava_m
 sft/smoe_refine_665k_llava/checkpoint-16632/logs/0409_1833_llava...mstar_llava_model_args_7bd459/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
 sft/smoe_refine_665k_llava/checkpoint-3328/logs/0409_1820_llava...mstar_llava_model_args_b94fe7/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
 sft/smoe_refine_665k_llava/checkpoint-3328/logs/0409_1820_llava...mstar_llava_model_args_b94fe7/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+sft/smoe_refine_665k_llava/logs/gqa/0412_1934_llava_v1.5_gqa_llava_model_args_4172a9/gqa.json filter=lfs diff=lfs merge=lfs -text
+sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/pope.json filter=lfs diff=lfs merge=lfs -text
+sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/textvqa_val.json filter=lfs diff=lfs merge=lfs -text

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/mmbench_en_dev.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:269f917d9576de83390b3278fb637ffe30f93762af58fb582620e85c50b24fd2
+size 14568181

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/rank0_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/rank1_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/rank2_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/rank3_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/results.json

@@ -0,0 +1,298 @@
+{
+    "results": {
+        "ai2d": {
+            "exact_match,flexible-extract": 0.6496113989637305,
+            "exact_match_stderr,flexible-extract": 0.008586842325753156,
+            "alias": "ai2d"
+        },
+        "mmbench_en_dev": {
+            "gpt_eval_score,none": 71.56357388316151,
+            "gpt_eval_score_stderr,none": "N/A",
+            "submission,none": null,
+            "submission_stderr,none": "N/A",
+            "alias": "mmbench_en_dev"
+        },
+        "mmmu_val": {
+            "mmmu_acc,none": 0.41889,
+            "mmmu_acc_stderr,none": "N/A",
+            "alias": "mmmu_val"
+        },
+        "mmstar": {
+            "coarse perception,none": 0.7266511799799377,
+            "coarse perception_stderr,none": "N/A",
+            "fine-grained perception,none": 0.3833578973225252,
+            "fine-grained perception_stderr,none": "N/A",
+            "instance reasoning,none": 0.5216438884434536,
+            "instance reasoning_stderr,none": "N/A",
+            "logical reasoning,none": 0.3982429012131982,
+            "logical reasoning_stderr,none": "N/A",
+            "math,none": 0.29996435890581835,
+            "math_stderr,none": "N/A",
+            "science & technology,none": 0.24473379748481924,
+            "science & technology_stderr,none": "N/A",
+            "alias": "mmstar"
+        }
+    },
+    "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 0x7f573be26af0>",
+            "doc_to_text": "<function ai2d_doc_to_text at 0x7f573be26dc0>",
+            "doc_to_target": "<function ai2d_doc_to_target at 0x7f573be2b280>",
+            "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 0x7f579c2d2af0>",
+            "doc_to_text": "<function mmbench_doc_to_text at 0x7f579c25d0d0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmbench_process_results at 0x7f579c25d670>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "gpt_eval_score",
+                    "aggregation": "<function mmbench_aggregate_dev_results_eval at 0x7f579c2d2280>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "submission",
+                    "aggregation": "<function mmbench_aggregate_dev_results_submission at 0x7f579c2d2820>",
+                    "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"
+                }
+            }
+        },
+        "mmmu_val": {
+            "task": "mmmu_val",
+            "dataset_path": "lmms-lab/MMMU",
+            "test_split": "validation",
+            "doc_to_visual": "<function mmmu_doc_to_visual at 0x7f574d84dc10>",
+            "doc_to_text": "<function mmmu_doc_to_text at 0x7f5773c2f8b0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmmu_process_results at 0x7f5773c45700>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mmmu_acc",
+                    "aggregation": "<function mmmu_aggregate_results at 0x7f57739225e0>",
+                    "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 0x7f573bd7dca0>",
+            "doc_to_text": "<function mmstar_doc_to_text at 0x7f573bd880d0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmstar_process_results at 0x7f573bd88550>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "coarse perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f573bd88940>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "fine-grained perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f573bd88ca0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "instance reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f573bd8e040>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "logical reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f573bd8e3a0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "science & technology",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f573bd8e700>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "math",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7f573bd8ea60>",
+                    "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"
+                }
+            }
+        }
+    },
+    "versions": {
+        "ai2d": "Yaml",
+        "mmbench_en_dev": "Yaml",
+        "mmmu_val": "Yaml",
+        "mmstar": "Yaml"
+    },
+    "n-shot": {
+        "ai2d": 0,
+        "mmbench_en_dev": 0,
+        "mmmu_val": 0,
+        "mmstar": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/sft/SMOE/test_Full_smoe,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "b61cb97"
+}

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/submissions/mmbench_en_dev_results.json

@@ -0,0 +1 @@
+{"overall_acc": 0.7156357388316151, "category_acc": {"action_recognition": 0.8888888888888888, "attribute_comparison": 0.75, "attribute_recognition": 0.8783783783783784, "celebrity_recognition": 0.8080808080808081, "function_reasoning": 0.8734177215189873, "future_prediction": 0.475, "identity_reasoning": 0.9777777777777777, "image_emotion": 0.84, "image_quality": 0.5283018867924528, "image_scene": 0.9519230769230769, "image_style": 0.8113207547169812, "image_topic": 0.8888888888888888, "nature_relation": 0.6875, "object_localization": 0.48148148148148145, "ocr": 0.6410256410256411, "physical_property_reasoning": 0.5866666666666667, "physical_relation": 0.4583333333333333, "social_relation": 0.7674418604651163, "spatial_relationship": 0.3111111111111111, "structuralized_imagetext_understanding": 0.41025641025641024}, "l2_category_acc": {"attribute_reasoning": 0.7889447236180904, "coarse_perception": 0.8243243243243243, "finegrained_perception (cross-instance)": 0.6643356643356644, "finegrained_perception (instance-level)": 0.7133105802047781, "logic_reasoning": 0.4322033898305085, "relation_reasoning": 0.6695652173913044}}

sft/smoe_refine_665k_llava/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0412_1911_llava...mstar_llava_model_args_4172a9/submissions/mmbench_en_dev_results.xlsx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b0a389ca00b157517d41368cff5f7ddefb68e34ec2d717f37f55baf31dab422
+size 865153

sft/smoe_refine_665k_llava/logs/gqa/0412_1934_llava_v1.5_gqa_llava_model_args_4172a9/gqa.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b3eb93ad63ce6a17250d269c1dea1f66b6bedbd6848e4c57bb8501c97901928
+size 38369383

sft/smoe_refine_665k_llava/logs/gqa/0412_1934_llava_v1.5_gqa_llava_model_args_4172a9/rank0_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/gqa/0412_1934_llava_v1.5_gqa_llava_model_args_4172a9/results.json

@@ -0,0 +1,79 @@
+{
+    "results": {
+        "gqa": {
+            "exact_match,none": 0.6148831292733344,
+            "exact_match_stderr,none": 0.004339143672998995,
+            "alias": "gqa"
+        }
+    },
+    "configs": {
+        "gqa": {
+            "task": "gqa",
+            "dataset_path": "lmms-lab/GQA",
+            "dataset_name": "testdev_balanced_instructions",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "testdev",
+            "doc_to_visual": "<function gqa_doc_to_visual at 0x7f337a112820>",
+            "doc_to_text": "<function gqa_doc_to_text at 0x7f337a063940>",
+            "doc_to_target": "answer",
+            "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,
+                "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."
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                }
+            }
+        }
+    },
+    "versions": {
+        "gqa": "Yaml"
+    },
+    "n-shot": {
+        "gqa": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/sft/SMOE/test_Full_smoe,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "b61cb97"
+}

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/mmerealworld_lite.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43d09dfabbd10426ed87e786b93da8fb445467154d1d8d354f9058a98445489f
+size 1994104230

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/pope.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04c67b7ca3ab3e41cc17c5f60c89d33c096a51fd55963a12f578a671cc905da5
+size 17577382

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/rank0_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/rank1_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/rank2_metric_eval_done.txt

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

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/results.json

@@ -0,0 +1,257 @@
+{
+    "results": {
+        "mmerealworld_lite": {
+            "mme_realworld_score,none": 0.32204273058884836,
+            "mme_realworld_score_stderr,none": "N/A",
+            "alias": "mmerealworld_lite"
+        },
+        "ocrbench": {
+            "ocrbench_accuracy,none": 0.323,
+            "ocrbench_accuracy_stderr,none": "N/A",
+            "alias": "ocrbench"
+        },
+        "pope": {
+            "pope_accuracy,none": 0.8651111111111112,
+            "pope_accuracy_stderr,none": "N/A",
+            "pope_precision,none": 0.9292058516196448,
+            "pope_precision_stderr,none": "N/A",
+            "pope_recall,none": 0.7904444444444444,
+            "pope_recall_stderr,none": "N/A",
+            "pope_f1_score,none": 0.8542267050912584,
+            "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": 0.4165600003123283,
+            "exact_match_stderr,none": 0.006747433492938556,
+            "submission,none": null,
+            "submission_stderr,none": "N/A",
+            "alias": "textvqa_val"
+        }
+    },
+    "configs": {
+        "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 0x7fa81e9ceaf0>",
+            "doc_to_text": "<function mme_realworld_doc_to_text at 0x7fa81e9d1700>",
+            "doc_to_target": "answer",
+            "process_results": "<function mme_realworld_process_results at 0x7fa81e9d74c0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mme_realworld_score",
+                    "aggregation": "<function mme_realworld_aggregate_results at 0x7fa81e9da160>",
+                    "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"
+                }
+            }
+        },
+        "ocrbench": {
+            "task": "ocrbench",
+            "dataset_path": "echo840/OCRBench",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function ocrbench_doc_to_visual at 0x7fa8337ce0d0>",
+            "doc_to_text": "<function ocrbench_doc_to_text at 0x7fa833117700>",
+            "doc_to_target": "answer",
+            "process_results": "<function ocrbench_process_results at 0x7fa833117a60>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "ocrbench_accuracy",
+                    "aggregation": "<function ocrbench_aggregate_accuracy at 0x7fa833117dc0>",
+                    "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
+                }
+            ]
+        },
+        "pope": {
+            "task": "pope",
+            "dataset_path": "lmms-lab/POPE",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function pope_doc_to_visual at 0x7fa81ea61e50>",
+            "doc_to_text": "<function pope_doc_to_text at 0x7fa81ea683a0>",
+            "doc_to_target": "answer",
+            "process_results": "<function pope_process_results at 0x7fa81ea688b0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "pope_accuracy",
+                    "aggregation": "<function pope_aggregate_accuracy at 0x7fa81ea68dc0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_precision",
+                    "aggregation": "<function pope_aggregate_precision at 0x7fa81ea69310>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_recall",
+                    "aggregation": "<function pope_aggregate_recall at 0x7fa81ea69820>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_f1_score",
+                    "aggregation": "<function pope_aggregate_f1_score at 0x7fa81ea69d30>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_yes_ratio",
+                    "aggregation": "<function pope_aggregate_yes_ratio at 0x7fa81ea71280>",
+                    "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 0x7fa84a8f41f0>",
+            "doc_to_text": "<function textvqa_doc_to_text at 0x7fa84a8f4670>",
+            "doc_to_target": "answer",
+            "process_results": "<function textvqa_process_results at 0x7fa84a8f4940>",
+            "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 0x7fa84a8f4040>",
+                    "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": {
+        "mmerealworld_lite": "Yaml",
+        "ocrbench": "Yaml",
+        "pope": "Yaml",
+        "textvqa_val": "Yaml"
+    },
+    "n-shot": {
+        "mmerealworld_lite": 0,
+        "ocrbench": 0,
+        "pope": 0,
+        "textvqa_val": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/sft/SMOE/test_Full_smoe,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "b61cb97"
+}

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/results/ocrbench_results.txt

@@ -0,0 +1,18 @@
+######################### OCRBench #############################
+Text Recognition(Total 300): 180
+---------------- Details of Recognition Score ------------------
+Regular Text Recognition(Total 50): 46
+Irregular Text Recognition(Total 50): 40
+Artistic Text Recognition(Total 50): 45
+Handwriting Recognition(Total 50): 28
+Digit String Recognition(Total 50): 9
+Non-Semantic Text Recognition(Total 50): 12
+----------------------------------------------------------------
+Scene Text-centric VQA(Total 200): 111
+----------------------------------------------------------------
+Doc-oriented VQA(Total 200): 23
+----------------------------------------------------------------
+Key Information Extraction(Total 200): 9
+Handwritten Mathematical Expression Recognition(Total 100): 0
+--------------------- Final Score ------------------------------
+Final Score(Total 1000): 323

sft/smoe_refine_665k_llava/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0412_1933_llava..._pope_llava_model_args_4172a9/textvqa_val.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb018759d20b23f05a8f6287cd9f52dbf3bd8915f6d5bffa4a679c0f4749b3bf
+size 13148492

sft/smoe_refine_665k_llava/model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5eed1597e630d33e971332c65f90e0abaa9a72d47769733b53e9bc39e1684d8f
+oid sha256:25d5a7cb279b9c52201e13da03fdcfcf3d4bf862f49543a29297a9c4bf94ca13
 size 4972489328

sft/smoe_refine_665k_llava/model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ddb3a759a0dd0e05fd3bc46899f02df24184e8eb1b117018b37fa3d6ef293dad
+oid sha256:ab05102ad84689c3a903f896d9e82da4045c9b8911256f625b22f4c5d8d99951
 size 4985529648

sft/smoe_refine_665k_llava/model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a43396d05647807918cc144c703116819da443aa175c98024702aa939391a252
+oid sha256:3044d9ec03437ea57f9f62abd7b681bbb31776b46339997b58a15ad2594db92a
 size 248943552

sft/smoe_refine_665k_llava/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1e89343be69cf9344a0c92b8a9b8d2f2e376d7e19a1cc7fe1cba698b6a65fa0c
-size 7928
+oid sha256:c8f88ca76a6d6eb76c2f7f274ca0c114e1f74df026b4738ee990a2acf8268c9b
+size 7992