Split (1)

train · 714k rows

text stringlengths 0 1.16k
torch.Size([7, 3, 448, 448])
torch.Size([7, 3, 448, 448])
question: ['How many keys 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']]
torch.Size([1, 3, 448, 448]) knan debug pixel values shape
tensor([0.2903, 0.2173, 0.1318, 0.0169, 0.0390, 0.0072, 0.2971, 0.0004],
device='cuda:2', grad_fn=<SoftmaxBackward0>)
2 *************
['1', '3', '4', '8', '6', '12', '2', '47'] tensor([0.2903, 0.2173, 0.1318, 0.0169, 0.0390, 0.0072, 0.2971, 0.0004],
device='cuda:2', grad_fn=<SelectBackward0>)
最后的概率分布为: {True: tensor(0.7097, device='cuda:2', grad_fn=<DivBackward0>), False: tensor(0.2903, device='cuda:2', grad_fn=<DivBackward0>), 'Execute Error': tensor(5.9605e-08, device='cuda:2', grad_fn=<DivBackward0>)}
question: ['What color are the vases?'], responses:['green']
question: ['How many bottles are in the image?'], responses:['6']
[('green', 0.1326115459908909), ('yellow', 0.12668030247077625), ('red', 0.12551779073733718), ('wild', 0.12324669870262604), ('orange and blue', 0.12319974118412196), ('bronze', 0.1230515752050065), ('pink', 0.12286305245049417), ('red white blue', 0.12282929325874692)]
[['green', 'yellow', 'red', 'wild', 'orange and blue', 'bronze', 'pink', 'red white blue']]
[('6', 0.12794147189263105), ('8', 0.12539492259598553), ('12', 0.12539359088927945), ('5', 0.12471292164321114), ('4', 0.12443617393590153), ('1', 0.12417386497855347), ('11', 0.12398049124372558), ('3', 0.12396656282071232)]
[['6', '8', '12', '5', '4', '1', '11', '3']]
torch.Size([7, 3, 448, 448]) knan debug pixel values shape
torch.Size([7, 3, 448, 448]) knan debug pixel values shape
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
tensor([8.6955e-01, 2.8468e-02, 9.7643e-02, 1.7639e-03, 1.8288e-04, 7.6360e-04,
9.2291e-05, 1.5321e-03], device='cuda:3', grad_fn=<SoftmaxBackward0>)
yes *************
['yes', 'congratulations', 'no', 'honey', 'solid', 'right', 'candle', 'chocolate'] tensor([8.6955e-01, 2.8468e-02, 9.7643e-02, 1.7639e-03, 1.8288e-04, 7.6360e-04,
9.2291e-05, 1.5321e-03], device='cuda:3', grad_fn=<SelectBackward0>)
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
最后的概率分布为: {True: tensor(0.8696, device='cuda:3', grad_fn=<DivBackward0>), False: tensor(0.0976, device='cuda:3', grad_fn=<DivBackward0>), 'Execute Error': tensor(0.0328, device='cuda:3', grad_fn=<DivBackward0>)}
ANSWER0=VQA(image=RIGHT,question='How many pillows are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} == 2')
FINAL_ANSWER=RESULT(var=ANSWER1)
torch.Size([5, 3, 448, 448])
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
dynamic ViT batch size: 7, images per sample: 7.0, dynamic token length: 1860
question: ['How many pillows are in the image?'], responses:['2']
tensor([9.3530e-01, 5.4443e-03, 6.6170e-03, 8.3545e-04, 7.3532e-03, 1.7705e-02,
2.4784e-02, 1.9599e-03], device='cuda:1', grad_fn=<SoftmaxBackward0>)
green *************
['green', 'yellow', 'red', 'wild', 'orange and blue', 'bronze', 'pink', 'red white blue'] tensor([9.3530e-01, 5.4443e-03, 6.6170e-03, 8.3545e-04, 7.3532e-03, 1.7705e-02,
2.4784e-02, 1.9599e-03], device='cuda:1', grad_fn=<SelectBackward0>)
最后的概率分布为: {True: tensor(0., device='cuda:1', grad_fn=<MulBackward0>), False: tensor(0., device='cuda:1', grad_fn=<MulBackward0>), 'Execute Error': tensor(1., device='cuda:1', grad_fn=<DivBackward0>)}
[('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']]
tensor([0.5150, 0.1894, 0.0094, 0.2422, 0.0240, 0.0019, 0.0145, 0.0036],
device='cuda:0', grad_fn=<SoftmaxBackward0>)
6 *************
['6', '8', '12', '5', '4', '1', '11', '3'] tensor([0.5150, 0.1894, 0.0094, 0.2422, 0.0240, 0.0019, 0.0145, 0.0036],
device='cuda:0', grad_fn=<SelectBackward0>)
最后的概率分布为: {True: tensor(0.5150, device='cuda:0', grad_fn=<DivBackward0>), False: tensor(0.4850, device='cuda:0', grad_fn=<DivBackward0>), 'Execute Error': tensor(5.9605e-08, device='cuda:0', grad_fn=<DivBackward0>)}
torch.Size([5, 3, 448, 448]) knan debug pixel values shape
tensor([7.6932e-01, 1.2558e-01, 3.3801e-02, 5.2349e-02, 1.2434e-02, 3.2429e-03,
3.0449e-03, 2.3042e-04], device='cuda:3', grad_fn=<SoftmaxBackward0>)
2 *************
['2', '3', '4', '1', '5', '8', '7', '29'] tensor([7.6932e-01, 1.2558e-01, 3.3801e-02, 5.2349e-02, 1.2434e-02, 3.2429e-03,
3.0449e-03, 2.3042e-04], device='cuda:3', grad_fn=<SelectBackward0>)
最后的概率分布为: {True: tensor(0.7693, device='cuda:3', grad_fn=<DivBackward0>), False: tensor(0.2307, device='cuda:3', grad_fn=<DivBackward0>), 'Execute Error': tensor(0., device='cuda:3', grad_fn=<DivBackward0>)}
[2024-10-23 14:47:59,001] [INFO] [logging.py:96:log_dist] [Rank 0] rank=0 time (ms) \| optimizer_allgather: 1.35 \| optimizer_gradients: 0.37 \| optimizer_step: 0.33
[2024-10-23 14:47:59,001] [INFO] [logging.py:96:log_dist] [Rank 0] rank=0 time (ms) \| forward_microstep: 5126.00 \| backward_microstep: 7515.82 \| backward_inner_microstep: 4836.01 \| backward_allreduce_microstep: 2679.74 \| step_microstep: 7.62
[2024-10-23 14:47:59,002] [INFO] [logging.py:96:log_dist] [Rank 0] rank=0 time (ms) \| forward: 5126.02 \| backward: 7515.81 \| backward_inner: 4836.02 \| backward_allreduce: 2679.70 \| step: 7.64

1%\| \| 27/4844 [06:42<17:42:32, 13.23s/it]Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
ANSWER0=VQA(image=RIGHT,question='Are triangular pennants on display in the image?')
ANSWER1=EVAL(expr='{ANSWER0}')
FINAL_ANSWER=RESULT(var=ANSWER1)
Registering VQA_lavis step
Registering EVAL step
Registering RESULT step
ANSWER0=VQA(image=RIGHT,question='How many virtually identical trifle desserts are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} == 2')
FINAL_ANSWER=RESULT(var=ANSWER1)
ANSWER0=VQA(image=LEFT,question='Is the animal's body turned to the right?')
ANSWER1=EVAL(expr='{ANSWER0}')
FINAL_ANSWER=RESULT(var=ANSWER1)
torch.Size([7, 3, 448, 448])
ANSWER0=VQA(image=LEFT,question='How many striped straws are in the image?')
ANSWER1=EVAL(expr='{ANSWER0} >= 1')
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 virtually identical trifle desserts are in the image?'], responses:['2']
question: ['Is the animal'], responses:['yes']
[('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']]
question: ['How many striped straws are in the image?'], responses:['0']
[('yes', 0.1298617250866936), ('congratulations', 0.12464161604141298), ('no', 0.12445222599225532), ('honey', 0.12437056445881921), ('solid', 0.12422595371654564), ('right', 0.12419889376311324), ('candle', 0.12414264780165109), ('chocolate', 0.12410637313950891)]
[['yes', 'congratulations', 'no', 'honey', 'solid', 'right', 'candle', 'chocolate']]
[('0', 0.13077743594303964), ('circles', 0.12449813349255197), ('maroon', 0.12428926693968681), ('large', 0.1242263466991631), ('rooster', 0.12409315512763705), ('nuts', 0.12408018414184876), ('beige', 0.1240288472550799), ('bottle', 0.12400663040099273)]

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.