dd1d6df5f4902c9b8cd21b3b129e2168962f1cd5bd125680152a8b6e5914d5e2
Browse files- sft/1M3/Full_smoe_tcmoe/logs/0629_0427_llava_v1.5_pope_llava_model_args_59313d/rank2_metric_eval_done.txt +1 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0427_llava_v1.5_pope_llava_model_args_59313d/rank3_metric_eval_done.txt +1 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0427_llava_v1.5_pope_llava_model_args_59313d/results.json +95 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/gpt_response/hallusion_output_vd_model.json +0 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/gpt_response/hallusion_output_vs_model.json +0 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/hallusion_bench_image.json +0 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank0_metric_eval_done.txt +1 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank1_metric_eval_done.txt +1 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank2_metric_eval_done.txt +1 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank3_metric_eval_done.txt +1 -0
- sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/results.json +87 -0
sft/1M3/Full_smoe_tcmoe/logs/0629_0427_llava_v1.5_pope_llava_model_args_59313d/rank2_metric_eval_done.txt
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
rank 2 eval done
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0427_llava_v1.5_pope_llava_model_args_59313d/rank3_metric_eval_done.txt
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
rank 3 eval done
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0427_llava_v1.5_pope_llava_model_args_59313d/results.json
ADDED
|
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"pope": {
|
| 4 |
+
"pope_accuracy,none": 0.8478888888888889,
|
| 5 |
+
"pope_accuracy_stderr,none": "N/A",
|
| 6 |
+
"pope_precision,none": 0.8689720065866855,
|
| 7 |
+
"pope_precision_stderr,none": "N/A",
|
| 8 |
+
"pope_recall,none": 0.8208888888888889,
|
| 9 |
+
"pope_recall_stderr,none": "N/A",
|
| 10 |
+
"pope_f1_score,none": 0.8442463718432178,
|
| 11 |
+
"pope_f1_score_stderr,none": "N/A",
|
| 12 |
+
"pope_yes_ratio,none": 0.5,
|
| 13 |
+
"pope_yes_ratio_stderr,none": "N/A",
|
| 14 |
+
"alias": "pope"
|
| 15 |
+
}
|
| 16 |
+
},
|
| 17 |
+
"configs": {
|
| 18 |
+
"pope": {
|
| 19 |
+
"task": "pope",
|
| 20 |
+
"dataset_path": "lmms-lab/POPE",
|
| 21 |
+
"dataset_kwargs": {
|
| 22 |
+
"token": true
|
| 23 |
+
},
|
| 24 |
+
"test_split": "test",
|
| 25 |
+
"doc_to_visual": "<function pope_doc_to_visual at 0x7feae07b48b0>",
|
| 26 |
+
"doc_to_text": "<function pope_doc_to_text at 0x7feae07b4e50>",
|
| 27 |
+
"doc_to_target": "answer",
|
| 28 |
+
"process_results": "<function pope_process_results at 0x7feae07bb430>",
|
| 29 |
+
"description": "",
|
| 30 |
+
"target_delimiter": " ",
|
| 31 |
+
"fewshot_delimiter": "\n\n",
|
| 32 |
+
"metric_list": [
|
| 33 |
+
{
|
| 34 |
+
"metric": "pope_accuracy",
|
| 35 |
+
"aggregation": "<function pope_aggregate_accuracy at 0x7feae07bb9d0>",
|
| 36 |
+
"higher_is_better": true
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"metric": "pope_precision",
|
| 40 |
+
"aggregation": "<function pope_aggregate_precision at 0x7feae07bbf70>",
|
| 41 |
+
"higher_is_better": true
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"metric": "pope_recall",
|
| 45 |
+
"aggregation": "<function pope_aggregate_recall at 0x7feae07bd550>",
|
| 46 |
+
"higher_is_better": true
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"metric": "pope_f1_score",
|
| 50 |
+
"aggregation": "<function pope_aggregate_f1_score at 0x7feae07bdaf0>",
|
| 51 |
+
"higher_is_better": true
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"metric": "pope_yes_ratio",
|
| 55 |
+
"aggregation": "<function pope_aggregate_yes_ratio at 0x7feae07bf0d0>",
|
| 56 |
+
"higher_is_better": true
|
| 57 |
+
}
|
| 58 |
+
],
|
| 59 |
+
"output_type": "generate_until",
|
| 60 |
+
"generation_kwargs": {
|
| 61 |
+
"max_new_tokens": 128,
|
| 62 |
+
"temperature": 0.0,
|
| 63 |
+
"top_p": 1.0,
|
| 64 |
+
"num_beams": 1,
|
| 65 |
+
"do_sample": false,
|
| 66 |
+
"until": [
|
| 67 |
+
"\n\n"
|
| 68 |
+
]
|
| 69 |
+
},
|
| 70 |
+
"repeats": 1,
|
| 71 |
+
"should_decontaminate": false,
|
| 72 |
+
"metadata": [
|
| 73 |
+
{
|
| 74 |
+
"version": 0.0
|
| 75 |
+
}
|
| 76 |
+
]
|
| 77 |
+
}
|
| 78 |
+
},
|
| 79 |
+
"versions": {
|
| 80 |
+
"pope": "Yaml"
|
| 81 |
+
},
|
| 82 |
+
"n-shot": {
|
| 83 |
+
"pope": 0
|
| 84 |
+
},
|
| 85 |
+
"model_configs": {
|
| 86 |
+
"model": "llava",
|
| 87 |
+
"model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/SMOE/1M3/Full_smoe_tcmoe,conv_template=phi35",
|
| 88 |
+
"batch_size": "1",
|
| 89 |
+
"device": null,
|
| 90 |
+
"limit": null,
|
| 91 |
+
"bootstrap_iters": 100000,
|
| 92 |
+
"gen_kwargs": ""
|
| 93 |
+
},
|
| 94 |
+
"git_hash": "289c7fe5"
|
| 95 |
+
}
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/gpt_response/hallusion_output_vd_model.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/gpt_response/hallusion_output_vs_model.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/hallusion_bench_image.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank0_metric_eval_done.txt
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
rank 0 eval done
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank1_metric_eval_done.txt
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
rank 1 eval done
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank2_metric_eval_done.txt
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
rank 2 eval done
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/rank3_metric_eval_done.txt
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
rank 3 eval done
|
sft/1M3/Full_smoe_tcmoe/logs/0629_0506_llava...image_llava_model_args_59313d/results.json
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"results": {
|
| 3 |
+
"hallusion_bench_image": {
|
| 4 |
+
"aAcc,none": 45.6362,
|
| 5 |
+
"aAcc_stderr,none": "N/A",
|
| 6 |
+
"fAcc,none": 21.3873,
|
| 7 |
+
"fAcc_stderr,none": "N/A",
|
| 8 |
+
"qAcc,none": 16.7033,
|
| 9 |
+
"qAcc_stderr,none": "N/A",
|
| 10 |
+
"alias": "hallusion_bench_image"
|
| 11 |
+
}
|
| 12 |
+
},
|
| 13 |
+
"configs": {
|
| 14 |
+
"hallusion_bench_image": {
|
| 15 |
+
"task": "hallusion_bench_image",
|
| 16 |
+
"dataset_path": "lmms-lab/HallusionBench",
|
| 17 |
+
"dataset_kwargs": {
|
| 18 |
+
"token": true
|
| 19 |
+
},
|
| 20 |
+
"test_split": "image",
|
| 21 |
+
"doc_to_visual": "<function hb_doc_to_visual at 0x7f63de76f280>",
|
| 22 |
+
"doc_to_text": "<function hb_doc_to_text at 0x7f63de76f9d0>",
|
| 23 |
+
"doc_to_target": "gt_answer_details",
|
| 24 |
+
"process_results": "<function hb_process_results at 0x7f63de78c1f0>",
|
| 25 |
+
"description": "",
|
| 26 |
+
"target_delimiter": " ",
|
| 27 |
+
"fewshot_delimiter": "\n\n",
|
| 28 |
+
"metric_list": [
|
| 29 |
+
{
|
| 30 |
+
"metric": "aAcc",
|
| 31 |
+
"aggregation": "<function hb_aggregation_result_aAcc at 0x7f63de78cb80>",
|
| 32 |
+
"higher_is_better": true
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"metric": "qAcc",
|
| 36 |
+
"aggregation": "<function hb_aggregation_result_qAcc at 0x7f63de69a1f0>",
|
| 37 |
+
"higher_is_better": true
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"metric": "fAcc",
|
| 41 |
+
"aggregation": "<function hb_aggregation_result_fAcc at 0x7f63de69a9d0>",
|
| 42 |
+
"higher_is_better": true
|
| 43 |
+
}
|
| 44 |
+
],
|
| 45 |
+
"output_type": "generate_until",
|
| 46 |
+
"generation_kwargs": {
|
| 47 |
+
"max_new_tokens": 128,
|
| 48 |
+
"temperature": 0.0,
|
| 49 |
+
"top_p": 1.0,
|
| 50 |
+
"num_beams": 1,
|
| 51 |
+
"do_sample": false,
|
| 52 |
+
"until": [
|
| 53 |
+
"\n\n"
|
| 54 |
+
]
|
| 55 |
+
},
|
| 56 |
+
"repeats": 1,
|
| 57 |
+
"should_decontaminate": false,
|
| 58 |
+
"metadata": [
|
| 59 |
+
{
|
| 60 |
+
"version": 0.0
|
| 61 |
+
}
|
| 62 |
+
],
|
| 63 |
+
"model_specific_prompt_kwargs": {
|
| 64 |
+
"default": {
|
| 65 |
+
"pre_prompt": "",
|
| 66 |
+
"post_prompt": ""
|
| 67 |
+
}
|
| 68 |
+
}
|
| 69 |
+
}
|
| 70 |
+
},
|
| 71 |
+
"versions": {
|
| 72 |
+
"hallusion_bench_image": "Yaml"
|
| 73 |
+
},
|
| 74 |
+
"n-shot": {
|
| 75 |
+
"hallusion_bench_image": 0
|
| 76 |
+
},
|
| 77 |
+
"model_configs": {
|
| 78 |
+
"model": "llava",
|
| 79 |
+
"model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/SMOE/1M3/Full_smoe_tcmoe,conv_template=phi35",
|
| 80 |
+
"batch_size": "1",
|
| 81 |
+
"device": null,
|
| 82 |
+
"limit": null,
|
| 83 |
+
"bootstrap_iters": 100000,
|
| 84 |
+
"gen_kwargs": ""
|
| 85 |
+
},
|
| 86 |
+
"git_hash": "289c7fe5"
|
| 87 |
+
}
|