Upload folder using huggingface_hub

.gitattributes

@@ -167,3 +167,9 @@ sft/1M3/Full_xmoe/logs/0626_1221_llava_v1.5_gqa_llava_model_args_2d4392/gqa.json
 sft/1M3/Full_xmoe/checkpoint-27572/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 sft/1M3/Full_xmoe/checkpoint-34462/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 sft/1M3/Full_xmoe/checkpoint-20679/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/pope.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+sft/1M3/std_0.002_Full_smoe/logs/0613_0314_llava_v1.5_gqa_llava_model_args_35cab6/gqa.json filter=lfs diff=lfs merge=lfs -text

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/mmbench_en_dev.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6f25b6a0beeedbbfb0c713dbea283fb787b953459cb8bbc6ac4292d8297c715
+size 14568190

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/mmerealworld_lite.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3734a62fb5df8bf60f2b0a9230e02fad8493220b461ed2b535505ca7a1174a53
+size 1994104223

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/pope.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c1da06ebc9c613d22568a13dd7d22b94426ad7b5055510c490e13c11d5c63bb
+size 17579727

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/rank0_metric_eval_done.txt

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

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/rank1_metric_eval_done.txt

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

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/rank2_metric_eval_done.txt

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

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/results.json

@@ -0,0 +1,488 @@
+{
+    "results": {
+        "ai2d": {
+            "exact_match,flexible-extract": 0.6910621761658031,
+            "exact_match_stderr,flexible-extract": 0.00831621652855342,
+            "alias": "ai2d"
+        },
+        "mmbench_en_dev": {
+            "gpt_eval_score,none": 70.79037800687286,
+            "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": 0.3397602918186555,
+            "mme_realworld_score_stderr,none": "N/A",
+            "alias": "mmerealworld_lite"
+        },
+        "mmmu_val": {
+            "mmmu_acc,none": 0.43444,
+            "mmmu_acc_stderr,none": "N/A",
+            "alias": "mmmu_val"
+        },
+        "mmstar": {
+            "coarse perception,none": 0.7033947521250198,
+            "coarse perception_stderr,none": "N/A",
+            "fine-grained perception,none": 0.35849914026037827,
+            "fine-grained perception_stderr,none": "N/A",
+            "instance reasoning,none": 0.5014961747473528,
+            "instance reasoning_stderr,none": "N/A",
+            "logical reasoning,none": 0.3802980298029803,
+            "logical reasoning_stderr,none": "N/A",
+            "math,none": 0.39419050164840064,
+            "math_stderr,none": "N/A",
+            "science & technology,none": 0.3278874261499387,
+            "science & technology_stderr,none": "N/A",
+            "alias": "mmstar"
+        },
+        "pope": {
+            "pope_accuracy,none": 0.8436666666666667,
+            "pope_accuracy_stderr,none": "N/A",
+            "pope_precision,none": 0.8854268891069677,
+            "pope_precision_stderr,none": "N/A",
+            "pope_recall,none": 0.8057602143335566,
+            "pope_recall_stderr,none": "N/A",
+            "pope_f1_score,none": 0.8437171244886031,
+            "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": 0.42702000027894976,
+            "exact_match_stderr,none": 0.006769072896119281,
+            "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 0x7fe3bc428280>",
+            "doc_to_text": "<function ai2d_doc_to_text at 0x7fe3bc428550>",
+            "doc_to_target": "<function ai2d_doc_to_target at 0x7fe3bc428af0>",
+            "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 0x7fe41c0c8af0>",
+            "doc_to_text": "<function mmbench_doc_to_text at 0x7fe41c056040>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmbench_process_results at 0x7fe41c056550>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "gpt_eval_score",
+                    "aggregation": "<function mmbench_aggregate_dev_results_eval at 0x7fe41c0c83a0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "submission",
+                    "aggregation": "<function mmbench_aggregate_dev_results_submission at 0x7fe41c0c88b0>",
+                    "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 0x7fe3bcbd0430>",
+            "doc_to_text": "<function mme_realworld_doc_to_text at 0x7fe3bcbd0ee0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mme_realworld_process_results at 0x7fe3bcbd6b80>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mme_realworld_score",
+                    "aggregation": "<function mme_realworld_aggregate_results at 0x7fe3bcbdd700>",
+                    "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 0x7fe3ced50e50>",
+            "doc_to_text": "<function mmmu_doc_to_text at 0x7fe3ced7cb80>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmmu_process_results at 0x7fe3cec5fa60>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mmmu_acc",
+                    "aggregation": "<function mmmu_aggregate_results at 0x7fe3cebb59d0>",
+                    "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 0x7fe3bc3838b0>",
+            "doc_to_text": "<function mmstar_doc_to_text at 0x7fe3bc383d30>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmstar_process_results at 0x7fe3bc389280>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "coarse perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fe3bc389700>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "fine-grained perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fe3bc389af0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "instance reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fe3bc389ee0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "logical reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fe3bc38f310>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "science & technology",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fe3bc38f700>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "math",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7fe3bc38faf0>",
+                    "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 0x7fe3bcc67ee0>",
+            "doc_to_text": "<function pope_doc_to_text at 0x7fe3bcc6b4c0>",
+            "doc_to_target": "answer",
+            "process_results": "<function pope_process_results at 0x7fe3bcc6ba60>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "pope_accuracy",
+                    "aggregation": "<function pope_aggregate_accuracy at 0x7fe3bcc6e040>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_precision",
+                    "aggregation": "<function pope_aggregate_precision at 0x7fe3bcc6e5e0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_recall",
+                    "aggregation": "<function pope_aggregate_recall at 0x7fe3bcc6eb80>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_f1_score",
+                    "aggregation": "<function pope_aggregate_f1_score at 0x7fe3bcc73160>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_yes_ratio",
+                    "aggregation": "<function pope_aggregate_yes_ratio at 0x7fe3bcc73700>",
+                    "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 0x7fe3e8d06ee0>",
+            "doc_to_text": "<function textvqa_doc_to_text at 0x7fe3e8c143a0>",
+            "doc_to_target": "answer",
+            "process_results": "<function textvqa_process_results at 0x7fe3e8c14670>",
+            "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 0x7fe3e8d06ca0>",
+                    "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/std_0.002_Full_smoe,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/submissions/mmbench_en_dev_results.json

@@ -0,0 +1 @@
+{"overall_acc": 0.7079037800687286, "category_acc": {"action_recognition": 0.8703703703703703, "attribute_comparison": 0.6590909090909091, "attribute_recognition": 0.9054054054054054, "celebrity_recognition": 0.7777777777777778, "function_reasoning": 0.8734177215189873, "future_prediction": 0.575, "identity_reasoning": 0.9777777777777777, "image_emotion": 0.82, "image_quality": 0.49056603773584906, "image_scene": 0.9519230769230769, "image_style": 0.8867924528301887, "image_topic": 0.8888888888888888, "nature_relation": 0.5833333333333334, "object_localization": 0.4074074074074074, "ocr": 0.6410256410256411, "physical_property_reasoning": 0.44, "physical_relation": 0.5833333333333334, "social_relation": 0.8372093023255814, "spatial_relationship": 0.2, "structuralized_imagetext_understanding": 0.5769230769230769}, "l2_category_acc": {"attribute_reasoning": 0.7336683417085427, "coarse_perception": 0.8277027027027027, "finegrained_perception (cross-instance)": 0.5944055944055944, "finegrained_perception (instance-level)": 0.689419795221843, "logic_reasoning": 0.576271186440678, "relation_reasoning": 0.6782608695652174}}

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/submissions/mmbench_en_dev_results.xlsx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:613d6406462bf5f7cdaa2cc19733a452f30570b8bce56bb4d51835a247d142c1
+size 843251

sft/1M3/std_0.002_Full_smoe/logs/0613_0312_llava..._pope_llava_model_args_35cab6/textvqa_val.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6be333dbcdf51b13d860f8881b1106e78c4fa9d3537a0f72b1a7adf5639b15d7
+size 13146060

sft/1M3/std_0.002_Full_smoe/logs/0613_0314_llava_v1.5_gqa_llava_model_args_35cab6/gqa.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c05f10a5e4ff3162c9290e5b2ad59725b82393898254c5599012ad1a2fc24946
+size 38379389

sft/1M3/std_0.002_Full_smoe/logs/0613_0314_llava_v1.5_gqa_llava_model_args_35cab6/rank0_metric_eval_done.txt

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

sft/1M3/std_0.002_Full_smoe/logs/0613_0314_llava_v1.5_gqa_llava_model_args_35cab6/results.json

@@ -0,0 +1,79 @@
+{
+    "results": {
+        "gqa": {
+            "exact_match,none": 0.4698680235331531,
+            "exact_match_stderr,none": 0.004450321794749798,
+            "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 0x7fdf7085ef70>",
+            "doc_to_text": "<function gqa_doc_to_text at 0x7fdf70737280>",
+            "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/SMOE/1M3/std_0.002_Full_smoe,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/std_0.002_Full_smoe/logs/0613_0336_llava...bench_llava_model_args_35cab6/rank0_metric_eval_done.txt

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

sft/1M3/std_0.002_Full_smoe/logs/0613_0336_llava...bench_llava_model_args_35cab6/rank1_metric_eval_done.txt

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

sft/1M3/std_0.002_Full_smoe/logs/0613_0336_llava...bench_llava_model_args_35cab6/rank2_metric_eval_done.txt

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

sft/1M3/std_0.002_Full_smoe/logs/0613_0336_llava...bench_llava_model_args_35cab6/results.json

@@ -0,0 +1,67 @@
+{
+    "results": {
+        "ocrbench": {
+            "ocrbench_accuracy,none": 0.402,
+            "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 0x7faafdac61f0>",
+            "doc_to_text": "<function ocrbench_doc_to_text at 0x7faafd4a2790>",
+            "doc_to_target": "answer",
+            "process_results": "<function ocrbench_process_results at 0x7faafd4a2af0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "ocrbench_accuracy",
+                    "aggregation": "<function ocrbench_aggregate_accuracy at 0x7faafd4a2e50>",
+                    "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/std_0.002_Full_smoe,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "289c7fe5"
+}

sft/1M3/std_0.002_Full_smoe/logs/0613_0336_llava...bench_llava_model_args_35cab6/results/ocrbench_results.txt

@@ -0,0 +1,18 @@
+######################### OCRBench #############################
+Text Recognition(Total 300): 215
+---------------- Details of Recognition Score ------------------
+Regular Text Recognition(Total 50): 47
+Irregular Text Recognition(Total 50): 44
+Artistic Text Recognition(Total 50): 43
+Handwriting Recognition(Total 50): 20
+Digit String Recognition(Total 50): 36
+Non-Semantic Text Recognition(Total 50): 25
+----------------------------------------------------------------
+Scene Text-centric VQA(Total 200): 112
+----------------------------------------------------------------
+Doc-oriented VQA(Total 200): 22
+----------------------------------------------------------------
+Key Information Extraction(Total 200): 18
+Handwritten Mathematical Expression Recognition(Total 100): 35
+--------------------- Final Score ------------------------------
+Final Score(Total 1000): 402