File size: 3,840 Bytes
d4bcd5c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 | === POPE done, TextVQA start воскресенье, 19 июля 2026 г. 00:21:10 (+07) ===
Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
Loading checkpoint shards: 0%| | 0/3 [00:00<?, ?it/s]
Loading checkpoint shards: 100%|██████████| 3/3 [00:00<00:00, 29.32it/s]
Loading checkpoint shards: 100%|██████████| 3/3 [00:00<00:00, 29.26it/s]
loaded llava-hf/llava-1.5-7b-hf on mps/torch.float16
TextVQA: 60 examples
10/60 608s (60.8s/ex)
20/60 889s (44.5s/ex)
30/60 1167s (38.9s/ex)
40/60 1469s (36.7s/ex)
50/60 1736s (34.7s/ex)
60/60 2070s (34.5s/ex)
{
"vanilla@576": {
"correct": 23.666666666666664,
"total": 60,
"accuracy": 39.44444444444444
},
"split@192": {
"correct": 19.666666666666664,
"total": 60,
"accuracy": 32.77777777777778
},
"split@128": {
"correct": 19.0,
"total": 60,
"accuracy": 31.666666666666668
},
"split@64": {
"correct": 18.0,
"total": 60,
"accuracy": 30.0
},
"random@192": {
"correct": 17.333333333333336,
"total": 60,
"accuracy": 28.888888888888893
},
"random@128": {
"correct": 13.333333333333332,
"total": 60,
"accuracy": 22.22222222222222
},
"random@64": {
"correct": 6.0,
"total": 60,
"accuracy": 10.0
},
"attn@192": {
"correct": 25.0,
"total": 60,
"accuracy": 41.666666666666664
},
"attn@128": {
"correct": 24.333333333333332,
"total": 60,
"accuracy": 40.55555555555555
},
"attn@64": {
"correct": 24.0,
"total": 60,
"accuracy": 40.0
}
}
wrote outputs/textvqa_results.json
=== TextVQA done, ScienceQA start воскресенье, 19 июля 2026 г. 00:56:15 (+07) ===
Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
Loading checkpoint shards: 0%| | 0/3 [00:00<?, ?it/s]
Loading checkpoint shards: 100%|██████████| 3/3 [00:00<00:00, 28.88it/s]
Loading checkpoint shards: 100%|██████████| 3/3 [00:00<00:00, 28.82it/s]
loaded llava-hf/llava-1.5-7b-hf on mps/torch.float16
ScienceQA-IMG: 60 examples
10/60 206s (20.6s/ex)
20/60 410s (20.5s/ex)
30/60 616s (20.5s/ex)
40/60 802s (20.1s/ex)
50/60 923s (18.5s/ex)
60/60 1044s (17.4s/ex)
{
"vanilla@576": {
"correct": 40.0,
"total": 60,
"accuracy": 66.66666666666667
},
"split@192": {
"correct": 39.0,
"total": 60,
"accuracy": 65.0
},
"split@128": {
"correct": 37.0,
"total": 60,
"accuracy": 61.666666666666664
},
"split@64": {
"correct": 37.0,
"total": 60,
"accuracy": 61.666666666666664
},
"random@192": {
"correct": 38.0,
"total": 60,
"accuracy": 63.333333333333336
},
"random@128": {
"correct": 38.0,
"total": 60,
"accuracy": 63.333333333333336
},
"random@64": {
"correct": 37.0,
"total": 60,
"accuracy": 61.666666666666664
},
"attn@192": {
"correct": 38.0,
"total": 60,
"accuracy": 63.333333333333336
},
"attn@128": {
"correct": 38.0,
"total": 60,
"accuracy": 63.333333333333336
},
"attn@64": {
"correct": 39.0,
"total": 60,
"accuracy": 65.0
}
}
wrote outputs/scienceqa_results.json
=== CHAIN2 DONE воскресенье, 19 июля 2026 г. 01:14:09 (+07) ===
|