text
stringlengths
0
1.16k
[2024-10-23 14:43:10,254] [INFO] [logging.py:96:log_dist] [Rank 0] rank=0 time (ms) | forward_microstep: 6431.59 | backward_microstep: 11386.35 | backward_inner_microstep: 6163.34 | backward_allreduce_microstep: 5222.90 | step_microstep: 7.51
[2024-10-23 14:43:10,255] [INFO] [logging.py:96:log_dist] [Rank 0] rank=0 time (ms) | forward: 6431.61 | backward: 11386.34 | backward_inner: 6163.36 | backward_allreduce: 5222.89 | step: 7.53
0%| | 7/4844 [01:54<21:10:32, 15.76s/it]Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
ANSWER0=VQA(image=LEFT,question='How many paper towel rolls are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} <= 1')
FINAL_ANSWER=RESULT(var=ANSWER1)
Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
ANSWER0=VQA(image=RIGHT,question='How many white capped bottles are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} >= 16')
FINAL_ANSWER=RESULT(var=ANSWER1)
Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
torch.Size([1, 3, 448, 448])
ANSWER0=VQA(image=RIGHT,question='How many pendant style lamps are above the bakery case in the image?')
ANSWER1=EVAL(expr='{ANSWER0} >= 2')
FINAL_ANSWER=RESULT(var=ANSWER1)
ANSWER0=VQA(image=LEFT,question='How many chimneys are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} >= 2')
FINAL_ANSWER=RESULT(var=ANSWER1)
torch.Size([7, 3, 448, 448])
torch.Size([13, 3, 448, 448])
torch.Size([7, 3, 448, 448])
question: ['How many white capped bottles are in the image?'], responses:['many']
[('many', 0.12680051474066337), ('few', 0.12559712123098582), ('several', 0.12545126119006317), ('blinds', 0.12452572291517987), ('moss', 0.12441899466837554), ('rainbow', 0.1244056457460399), ('kite', 0.12440323404357946), ('directions', 0.12439750546511286)]
[['many', 'few', 'several', 'blinds', 'moss', 'rainbow', 'kite', 'directions']]
torch.Size([1, 3, 448, 448]) knan debug pixel values shape
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 326
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 327
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 327
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 327
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 327
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 327
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 327
dynamic ViT batch size: 1, images per sample: 1.0, dynamic token length: 327
tensor([0.6048, 0.1314, 0.1858, 0.0088, 0.0135, 0.0314, 0.0049, 0.0195],
device='cuda:0', grad_fn=<SoftmaxBackward0>)
many *************
['many', 'few', 'several', 'blinds', 'moss', 'rainbow', 'kite', 'directions'] tensor([0.6048, 0.1314, 0.1858, 0.0088, 0.0135, 0.0314, 0.0049, 0.0195],
device='cuda:0', grad_fn=<SelectBackward0>)
最后的概率分布为: {True: tensor(0., device='cuda:0', grad_fn=<MulBackward0>), False: tensor(0., device='cuda:0', grad_fn=<MulBackward0>), 'Execute Error': tensor(1., device='cuda:0', grad_fn=<DivBackward0>)}
ANSWER0=VQA(image=LEFT,question='How many white dogs are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} >= 2')
FINAL_ANSWER=RESULT(var=ANSWER1)
torch.Size([7, 3, 448, 448])
question: ['How many pendant style lamps are above the bakery case in the image?'], responses:['2']
question: ['How many chimneys are in the image?'], responses:['1']
[('2', 0.12961991198727602), ('3', 0.12561270547489775), ('4', 0.12556127085987287), ('1', 0.1254920833223361), ('5', 0.12407835939022728), ('8', 0.124024076973589), ('7', 0.12288810153923228), ('29', 0.12272349045256851)]
[['2', '3', '4', '1', '5', '8', '7', '29']]
[('1', 0.12829009354978346), ('3', 0.12529928082343206), ('4', 0.12464806219229535), ('8', 0.12460015878893425), ('6', 0.12451220062887247), ('12', 0.124338487048427), ('2', 0.12420459433498025), ('47', 0.12410712263327517)]
[['1', '3', '4', '8', '6', '12', '2', '47']]
torch.Size([7, 3, 448, 448]) knan debug pixel values shape
torch.Size([7, 3, 448, 448]) knan debug pixel values shape
question: ['How many paper towel rolls are in the image?'], responses:['1']
question: ['How many white dogs are in the image?'], responses:['1']
[('1', 0.12829009354978346), ('3', 0.12529928082343206), ('4', 0.12464806219229535), ('8', 0.12460015878893425), ('6', 0.12451220062887247), ('12', 0.124338487048427), ('2', 0.12420459433498025), ('47', 0.12410712263327517)]
[['1', '3', '4', '8', '6', '12', '2', '47']]
[('1', 0.12829009354978346), ('3', 0.12529928082343206), ('4', 0.12464806219229535), ('8', 0.12460015878893425), ('6', 0.12451220062887247), ('12', 0.124338487048427), ('2', 0.12420459433498025), ('47', 0.12410712263327517)]
[['1', '3', '4', '8', '6', '12', '2', '47']]
torch.Size([7, 3, 448, 448]) knan debug pixel values shape
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
torch.Size([13, 3, 448, 448]) knan debug pixel values shape
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
tensor([4.2009e-01, 2.9054e-01, 1.3704e-01, 7.8195e-02, 5.0414e-02, 1.0257e-02,
1.3152e-02, 3.1381e-04], device='cuda:1', grad_fn=<SoftmaxBackward0>)
2 *************
['2', '3', '4', '1', '5', '8', '7', '29'] tensor([4.2009e-01, 2.9054e-01, 1.3704e-01, 7.8195e-02, 5.0414e-02, 1.0257e-02,
1.3152e-02, 3.1381e-04], device='cuda:1', grad_fn=<SelectBackward0>)
最后的概率分布为: {True: tensor(0.9218, device='cuda:1', grad_fn=<DivBackward0>), False: tensor(0.0782, device='cuda:1', grad_fn=<DivBackward0>), 'Execute Error': tensor(5.9605e-08, device='cuda:1', grad_fn=<DivBackward0>)}
tensor([6.1156e-01, 6.8598e-02, 2.0921e-02, 2.6596e-03, 5.8091e-03, 1.4955e-03,
2.8888e-01, 7.8209e-05], device='cuda:2', grad_fn=<SoftmaxBackward0>)
1 *************
['1', '3', '4', '8', '6', '12', '2', '47'] tensor([6.1156e-01, 6.8598e-02, 2.0921e-02, 2.6596e-03, 5.8091e-03, 1.4955e-03,
2.8888e-01, 7.8209e-05], device='cuda:2', grad_fn=<SelectBackward0>)
ANSWER0=VQA(image=LEFT,question='How many dogs are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} == 1')
FINAL_ANSWER=RESULT(var=ANSWER1)
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
最后的概率分布为: {True: tensor(0.3884, device='cuda:2', grad_fn=<DivBackward0>), False: tensor(0.6116, device='cuda:2', grad_fn=<DivBackward0>), 'Execute Error': tensor(-1.1921e-07, device='cuda:2', grad_fn=<DivBackward0>)}
torch.Size([7, 3, 448, 448])
ANSWER0=VQA(image=LEFT,question='How many animals are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} == 2')
FINAL_ANSWER=RESULT(var=ANSWER1)
torch.Size([13, 3, 448, 448])
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1861
tensor([8.7571e-01, 2.6455e-02, 1.3710e-02, 5.2020e-03, 7.1169e-03, 3.9104e-03,
6.7488e-02, 4.1040e-04], device='cuda:0', grad_fn=<SoftmaxBackward0>)
1 *************